DETAILED ACTION
Claims 1-17 and 18 and 20 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (math and mental process and human activity) without significantly more:
Claim(s) 1,15,16,17 and 18 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1).
Claim(s) 2 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Tu et al. (US 2023/0070286 A1):
Claim(s) 3,4 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Tu et al. (US 2023/0070286 A1) as applied in claim 2 further in view of Martinez Manzano et al. (US 2023/0401590 A1):
Claim(s) 5,9 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1):
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Xiao et al. (US 2023/0274248 A1):
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Bassi (US 2006/0050074 A1):
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Bassi (US 2006/0050074 A1) as applied in claim 7 further in view of OGASAWARA et al. (US 2023/0316489 A1):
Claim(s) 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied to claims 5,9 further in view of Hossain et al. (Bi-SAN-CAP: Bi-Directional Self-Attention for Image Captioning):
Claim(s) 11,12 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied to claims 5,9 further in view of LI et al. (CN 113269724 A) with SEARCH machine translation:
Claim(s) 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of KAZUHIRO (JP 2016-158059 A) with SEARCH machine translation:
Claim(s) 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of KAZUHIRO (JP 2016-158059 A) with SEARCH machine translation as applied in claim 13 further in view of Jaiswal et al. (US 11,775,617 B1):
Claim(s) 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17,18,19 further in view of Chiu et al. (US 2020/0357143 A1) further in view of LIU et al. (CN 112163608 A) with SEARCH machine translation:
Response to Amendment
The preliminary amendment was received 4/25/2024. Claims 1-20 pending:
PNG
media_image1.png
724
156
media_image1.png
Greyscale
Priority
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, 365(c), or 386(c) is acknowledged. Applicant has not complied with one or more conditions for receiving the benefit of an earlier filing date under 35 U.S.C. 120, 121, 365(c), or 386(c) as follows:
The later-filed application must be an application for a patent for an invention which is also disclosed in the prior application (the parent or original nonprovisional application or provisional application). The disclosure of the invention in the parent application and in the later-filed application must be sufficient to comply with the requirements of 35 U.S.C. 112(a) or the first paragraph of pre-AIA 35 U.S.C. 112, except for the best mode requirement. See Transco Products, Inc. v. Performance Contracting, Inc., 38 F.3d 551, 32 USPQ2d 1077 (Fed. Cir. 1994).
The disclosure of the prior-filed application, Application No. PRO 63/253,514 10/07/2021, fails to provide adequate support or enablement in the manner provided by 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph for one or more claims of this application:
The claimed “histological feature” of claims 1,18,19,20 is not in Application No. PRO 63/253,514 10/07/2021.
The claimed “the spatial attention models attention to a microscopic visual pattern” of claims 20 is not in Application No. PRO 63/253,514 10/07/2021.
The claimed “the sematic attention models attention to a macroscopic visual pattern” of claims 20 is not in Application No. PRO 63/253,514 10/07/2021.
Accordingly, claims 1-20 are not entitled to the benefit of the prior application (Application No. PRO 63/253,514 10/07/2021).
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-17 and 18 and 20 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (math and mental process and human activity) without significantly more:
PNG
media_image1.png
724
156
media_image1.png
Greyscale
Step Zero: Establish Broadest Reasonable Interpretation (in footnotes throughout this Office action);
Step 1: Claim 1 is a process; Claim 18 is manufacture; clam 19 a machine; claim 20 a process.
Step 2A, prong 1:
The claim(s) recite(s) the abstract: “extracting1 an embedding”2; “the embedding” “encoding3 the corresponding embedding” “representation4” “combining5 the encoded embeddings” “performing a pathological task”6 “).
1. (Currently Amended) A computer-implemented method for analyzing a whole slide image (WSI) in light of biological context, comprising:
extracting an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches of the set of patches, generating7 an encoded embedding for the patch by:
incorporating into the extracted embedding wherein the local pattern is derived from embeddings of one or more neighboring patches; and
incorporating into the extracted embedding related to the one or more histological features, wherein the global pattern is derived from embeddings of
generating a representation for the WSI by combining the encoded
performing a pathological task based on the representation for the WSI.
Step 2A, prong 2:
This judicial exception is not integrated into a practical application because the additional elements (“patches…WSI…histological features…local pattern…a global patten”) does not improve technology or technical field of the functioning of a computer or reflect thereof in view of applicant’s disclosure (paragraphs [2][33][34][35][59]).
1. (Currently Amended) A computer-implemented method for analyzing a whole slide image (WSI) in light of biological context, comprising:
extracting an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches of the set of patches, generating8 an encoded embedding for the patch by:
incorporating into the extracted embedding a local pattern related to the one or more histological features, wherein the local pattern is derived from embeddings of one or more neighboring patches; and
incorporating into the extracted embedding a global pattern related to the one or more histological features, wherein the global pattern is derived from embeddings of
generating a representation for the WSI by combining the encoded
performing a pathological task based on the representation for the WSI.
In contrast, claim 19 is recognized as machine learning9 “as providing the improvement” and thus “reflects the disclosed improvement” (MPEP 2106.04(d)(1), 2nd para) in machine learning as compared to the prior teachings of machine learning (i.e., multi-instance learning) in the disclosure at [65], last S: “combination”:
PNG
media_image2.png
1120
1009
media_image2.png
Greyscale
Step 2B:
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements adhere to conventional practices in view of applicant’s disclosure (BACKGROUND [3]):
PNG
media_image3.png
670
916
media_image3.png
Greyscale
1. (SUGGESTED1011) A computer-implemented method for analyzing a whole slide image (WSI) in light of biological context, comprising:
extracting an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches of the set of patches, generating [[,]] by a neural network model [[,]]12 an encoded embedding for the patch by:
incorporating into the extracted embedding wherein the local pattern is derived from embeddings of one or more neighboring patches; and
incorporating into the extracted embedding related to the one or more histological features, wherein the global pattern is derived from embeddings of
generating a representation for the WSI by combining the encoded
performing a pathological task based on the representation for the WSI.
Response to Arguments
Applicant's arguments filed 4/27/2026, pages 10-29 have been fully considered but they are not persuasive.
III. Priority
Applicant’s state on page 10:
The Office Action indicates the claimed "global pattern" of claims 1, 18, 19 is allegedly not supported in the provisional application. Applicant respectfully disagrees. For example, paragraph [14] of the provisional application discloses "(i) a semantic self-attention to the appearance of all other patches in the slide to model slide-level patterns as context (e.g., macroscopic context), and (ii) a spatial self-attention to nearby patches to disambiguate region-level patterns (e.g., microscopic context)." (Provisional Application, [14]; see also Id., [18].) Given the disclosure of the current application, a person having ordinary skill in the art would understand that "global pattern" can correspond to these slide-level (or macroscopic) patterns:
[3] This contrasts with diagnostic practice where a pathologist refers to both microscopic (regional or local within the WSI) patterns and macroscopic (global within the WSI) context when analyzing WSis.
[4] The transformer-based aggregation model may encode the embedding for each patch with two types of self-attention: a semantic self-attention to the appearance of all other patches in the slide to model slide-level patterns as global context (macroscopic context) and a spatial self-attention to nearby patches to disambiguate local patterns (microscopic context).
[35] Specifically, the transformer-based aggregation model includes two types of self-attention for each patch: (i) a semantic self-attention, which combines information about the appearance of all other patches in the slide to model slide-level patterns as global context (e.g., macroscopic context), and (ii) a spatial self-attention, which combines information about nearby patches to disambiguate local patterns (e.g., microscopic context).
In response “ ("'consideration of the understanding of one skilled in the art in no way relieves the patentee of adequately disclosing sufficient structure in the specification.’ It is not enough for the patentee simply to state or later argue that persons of ordinary skill in the art would know what structures to use to accomplish the claimed function."), quoting Atmel Corp. v. Information Storage Devices, Inc., 198 F.3d 1374, 1380, 53 USPQ2d 1225, 1229 (Fed. Cir. 1999); Biomedino, LLC v. Waters Technologies Corp., 490 F.3d 946, 953, 83 USPQ2d 1118, 1123 (Fed. Cir. 2007) ("The inquiry is whether one of skill in the art would understand the specification itself to disclose a structure, not simply whether that person would be capable of implementing a structure.").” via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(a) Original claims, 7th txt blk:
If a claim limitation invokes 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it must be interpreted to cover the corresponding structure, materials, or acts in the specification and "equivalents thereof." See 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. See also B. Braun Medical, Inc. v. Abbott Labs., 124 F.3d 1419, 1424, 43 USPQ2d 1896, 1899 (Fed. Cir. 1997). In considering whether there is 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, support for a means- (or step) plus- function claim limitation, the examiner must consider not only the original disclosure contained in the summary and detailed description of the invention portions of the specification, but also the original claims, abstract, and drawings. A means- (or step-) plus- function claim limitation is adequately described under 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, if: (1) The written description adequately links or associates adequately described particular structure, material, or acts to perform the function recited in a means- (or step-) plus- function claim limitation; or (2) it is clear based on the facts of the application that one skilled in the art would have known what structure, material, or acts disclosed in the specification perform the function recited in a means- (or step-) plus- function limitation. See Aristocrat Techs. Australia PTY Ltd. v. Int’l Game Tech., 521 F.3d 1328, 1336-37, 86 USPQ2d 1235, 1242 (Fed. Cir. 2008) ("'consideration of the understanding of one skilled in the art in no way relieves the patentee of adequately disclosing sufficient structure in the specification.’ It is not enough for the patentee simply to state or later argue that persons of ordinary skill in the art would know what structures to use to accomplish the claimed function."), quoting Atmel Corp. v. Information Storage Devices, Inc., 198 F.3d 1374, 1380, 53 USPQ2d 1225, 1229 (Fed. Cir. 1999); Biomedino, LLC v. Waters Technologies Corp., 490 F.3d 946, 953, 83 USPQ2d 1118, 1123 (Fed. Cir. 2007) ("The inquiry is whether one of skill in the art would understand the specification itself to disclose a structure, not simply whether that person would be capable of implementing a structure."). Note also that a rejection under 35 U.S.C. 112(b) or pre-AIA 35 U.S.C. 112, second paragraph, "cannot stand where there is adequate description in the specification to satisfy 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, regarding means-plus-function recitations that are not, per se, challenged for being unclear." In re Noll, 545 F.2d 141, 149, 191 USPQ 721, 727 (CCPA 1976). See "Supplemental Examination Guidelines for Determining the Applicability of 35 U.S.C. 112, para. 6," 65 Fed. Reg. 38510, June 21, 2000; see also MPEP § 2181. However, when a means- (or step-) plus-function claim limitation is found to be indefinite based on failure of the specification to disclose sufficient corresponding structure, materials, or acts that perform the entire claimed function, then the claim limitation necessarily lacks an adequate written description. Thus, when a claim is rejected as indefinite under 35 U.S.C. 112(b) or pre-AIA 35 U.S.C. 112, second paragraph because there is no corresponding structure, materials, or acts, or an inadequate disclosure of corresponding structure, materials, or acts, for a means- (or step-) plus-function claim limitation, then the claim must also be rejected under 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, for lack of an adequate written description.
Upon further review of 63/253,514, the claimed “global13 pattern” of claim 1 is supported in 63/253,514 or means the same as “macroscopic…pattern” via 63/253,514, paragraph [18]:
[18] Semantic and Spatial Self-Attention: Semantic and spatial self-attentions are explicitly encoded with two separate types of encoders. The semantic encoding (FIG. 1, [b]) models macroscopic14 visual patterns: it correlates a patch P with the whole slide by attending embeddings of all other patches to P as semantic context. This is motivated by the requirement of clinical diagnosis, where the impact of local sub-cellular patterns may depend on the co-existence of other patterns in a slide. Such semantic dependence can be achieved by using a bidirectional self-attention encoder with multi-head, with explicit cross-attention of patch) to patch i denoted
as a-.
ij
[
Applicants state in page 11:
The Office Action indicates the claimed "histological feature" of claims 1, 18, 19, 20 is allegedly not supported in the provisional application. Applicant respectfully disagrees. The provisional application recites numerous histological features, for example:
[50] [A] tile embedding can include one or more features that indicate and/or correspond to a size of depicted objects (e.g., sizes of depicted cells or aberrations) and/or density of depicted objects (e.g., a density of depicted cells or aberrations).
[18] This is motivated by the requirement of clinical diagnosis, where the impact of local sub-cellular patterns may depend on the co-existence of other patterns in a slide."
[19] Since sub-cellular structures can be of different scales in WSis, the spatial encoding models the regional visual patterns that extend beyond the scope of a single patch.
A person having ordinary skill in the art would understand that such features indicating or corresponding to the sizes of cells, densities of cells, and sub-cellular structures are histological features. For example, Dictionary.com defines "histological" as "of or relating to organic tissues or their structure." (https://www.dictionary.com/browse/histological.) Thus, Applicant respectfully submits the term "histological feature" is supported by the provisional
application.
In response “ ("'consideration of the understanding of one skilled in the art in no way relieves the patentee of adequately disclosing sufficient structure in the specification.’ It is not enough for the patentee simply to state or later argue that persons of ordinary skill in the art would know what structures to use to accomplish the claimed function."), quoting Atmel Corp. v. Information Storage Devices, Inc., 198 F.3d 1374, 1380, 53 USPQ2d 1225, 1229 (Fed. Cir. 1999); Biomedino, LLC v. Waters Technologies Corp., 490 F.3d 946, 953, 83 USPQ2d 1118, 1123 (Fed. Cir. 2007) ("The inquiry is whether one of skill in the art would understand the specification itself to disclose a structure, not simply whether that person would be capable of implementing a structure.").” via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(a) Original claims, 7th txt blk.
Upon further review of 63/253,514 and above said paragraph [50], the claimed "histological15 feature" is not recognized/spotted/identified in 63/253,514 at said [50]. Instead, 63/253,514’s paragraph [50] discloses “one or more features that indicate and/or correspond to a size of depicted objects (e.g., sizes of depicted cells16 or aberrations17) and/or density of depicted objects (e.g., a density of depicted cells or aberrations)” via [50]
[50] As described herein, tile embeddings can be generated from a deep learning neural
network using visual features of the tiles. Tile embeddings can be further generated from
contextual information associated with the tiles or from the content shown in the tile. For example,
a tile embedding can include one or more features that indicate and/or correspond to a size of
depicted objects (e.g., sizes of depicted cells or aberrations) and/or density of depicted objects
(e.g., a density of depicted cells or aberrations). Size and density can be measured absolutely (e.g.,
width expressed in pixels or converted from pixels to nanometers) or relative to other tiles from
the same digital pathology image, from a class of digital pathology images ( e.g., produced using
similar techniques or by a single whole slide image generation system or scanner), or from a related
family of digital pathology images. Furthermore, tiles can be classified prior to the tile embedding
module 312 generating embeddings for the tiles such that the tile embedding module 312 considers
the classification when preparing the embeddings. In particular embodiments, tile embedding
module 312 may incorporate one or more aspects of the WSITrans model.
Thus 63/253,251 discloses a cell-feature and an aberration-feature; however, claim 1’s “histological feature” “is not apparent” in 63/253,251 “and the examiner finds that the disclosure does not reasonably convey that the inventor had possession of the subject matter” via MPEP: 2163 II. A., 2nd para:
MPEP 2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, 2nd para:
With respect to newly added or amended claims, applicant should show support in the original disclosure for the new or amended claims. See, e.g., Hyatt v. Dudas, 492 F.3d 1365, 1370, n.4, 83 USPQ2d 1373, 1376, n.4 (Fed. Cir. 2007) (citing MPEP § 2163.04 which provides that a "simple statement such as ‘applicant has not pointed out where the new (or amended) claim is supported, nor does there appear to be a written description of the claim limitation ‘___’ in the application as filed’ may be sufficient where the claim is a new or amended claim, the support for the limitation is not apparent, and applicant has not pointed out where the limitation is supported."); see also MPEP §§ 714.02 and 2163.06 ("Applicant should ... specifically point out the support for any amendments made to the disclosure."); and MPEP § 2163.04 ("If applicant amends the claims and points out where and/or how the originally filed disclosure supports the amendment(s), and the examiner finds that the disclosure does not reasonably convey that the inventor had possession of the subject matter of the amendment at the time of the filing of the application, the examiner has the initial burden of presenting evidence [63/253,251’s paragraphs [34] that discloses “histological”-“sample” & [50] that discloses various object features such as cell-feature] or reasoning (histological is relating to organic tissues or their structures, while the disclosed “cell” and “aberration” of said [50] are not recognized in said 63/253,251 as relating to organic tissues or their structures, for example an aberration -a term in biology- can be a mode of organization and is silent regarding relating to organic tissues or their structures; thus, one of skill in the art would have to decipher18 “histological feature” from the disclosure of 63/253,251’s (1) “aberration”-“feature”, being silent regarding relating to organic tissues or their structure, and (2) “histological”-“sample”) to explain why persons skilled in the art would not recognize in the disclosure a description of the invention defined by the claims."). The inquiry into whether the description requirement is met is a question of fact that must be determined on a case-by-case basis. AbbVie Deutschland GmbH & Co., KG v. Janssen Biotech, Inc., 759 F.3d 1285, 1297, 111 USPQ2d 1780, 1788 (Fed. Cir. 2014) ("Whether a patent claim is supported by an adequate written description is a question of fact."); In re Smith, 458 F.2d 1389, 1395, 173 USPQ 679, 683 (CCPA 1972) ("Precisely how close [to the claimed invention] the description must come to comply with Sec. 112 must be left to case-by-case development."); In re Wertheim, 541 F.2d at 262, 191 USPQ at 96 (inquiry is primarily factual and depends on the nature of the invention and the amount of knowledge imparted to those skilled in the art by the disclosure).
Upon further review of 63/253,514 (filed 10/07/2021), “the level19 of…knowledge in the art” as of 29 AUG 2022 as discussed in Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images) is fail20-to-achieve or falls-short-to-achieve incorporation of global, feature information21: Wang, pg. 3967, lcol, 2nd para, penult S::
Moreover, pathologists usually make use of both local texture information and global information (e.g., shape, location) to make decisions, while our proposed approach, as well as other patch-based methods, merely depend on the local
texture and fail to incorporate global information.
and the claimed “histological22 feature23” is also not apparent in the “knowledge in the art” of said Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images: I don’t see the adjective “histological” (or forms24 thereof such as histopathology25) modifying26 the noun “feature” (or forms thereof such as class27 : histopathology class=histological feature) in Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images)) via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(a) Original claims
i) For Each Claim Drawn to a Single Embodiment or Species, 6th txt blk:
Whether the specification [63/253,514] shows that the inventor was in possession of the claimed invention (claim 1’s “histological feature”) is not a single, simple determination, but rather is a factual determination reached by considering a number of factors. Factors to be considered in determining whether there is sufficient evidence of possession include the level of skill and knowledge in the art [Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images)], partial structure, physical and/or chemical properties, functional characteristics alone or coupled with a known or disclosed correlation between structure and function, and the method of making the claimed invention. Disclosure of any combination of such identifying characteristics that distinguish the claimed invention from other materials and would lead one of skill in the art to the conclusion that the inventor was in possession of the claimed species is sufficient. See Eli Lilly, 119 F.3d at 1568, 43 USPQ2d at 1406. The description needed to satisfy the requirements of 35 U.S.C. 112 "varies with the nature and scope of the invention at issue, and with the scientific and technologic knowledge [i.e., Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images)] already in existence." Capon v. Eshhar, 418 F.3d at 1357, 76 USPQ2d at 1084. Patents and printed publications [Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images): printed 29 AUG 2022] in the art should be relied upon to determine whether an art is mature and what the level [“fail”: Wang, pg. 3967, lcol, 2nd para, penult S] of knowledge and skill is in the art. In most technologies which are mature, and wherein the knowledge and level of skill in the art is high, a written description question should not be raised for claims present in the application [63/253,514] when originally filed [10/07/2021], even if the specification discloses only a method of making the invention and the function of the invention.
Thus the above knowledge in the art (Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images)) is one of “a number of factors” to determine whether the inventor was in possession and full claim scope thereof of the claimed invention (claim 1’s “histological feature”).
Applicant’s disclosure of said “sub28-cellular29 patterns30”:
[18] This is motivated by the requirement of clinical diagnosis, where the impact of local sub-cellular patterns may depend on the co-existence of other patterns in a slide."
[19] Since sub-cellular structures can be of different scales in WSis, the spatial encoding models the regional visual patterns that extend beyond the scope of a single patch.
is directed to:
underneath the cell samples or
a secondary part of the cell samples or
less than a complete cell samples
which is less in scope to histological31:
PNG
media_image4.png
585
760
media_image4.png
Greyscale
Applicant’s disclosure of said “sub32-cellular33 structures34”:
[18] This is motivated by the requirement of clinical diagnosis, where the impact of local sub-cellular patterns may depend on the co-existence of other patterns in a slide."
[19] Since sub-cellular structures can be of different scales in WSis, the spatial encoding models the regional visual patterns that extend beyond the scope of a single patch.
is directed to:
underneath the cell construction and arrangement of tissues or
a secondary part of the cell construction and arrangement of tissues or
less than a complete cell construction and arrangement of tissues
which is also less in scope to histological353637:
PNG
media_image4.png
585
760
media_image4.png
Greyscale
Thus the claimed “histological feature” is not apparent38 or is not plain or clear in 63/253,514
Applicants state in pages 11,12:
The Office Action indicates the claimed "the spatial attention models attention to a microscopic visual pattern" of claim 20 is allegedly not supported in the provisional application. Applicant respectfully disagrees. The provisional application describes:
[14] In the present embodiments, a transformer-based aggregation model ("WSITrans") models cross-patch dependencies between all patches to capture microscopic and macroscopic patterns in WSis. Specifically, WSITrans includes two types of self-attention for each patch: (i) a semantic self-attention to the appearance of all other patches in the slide to model slide-level patterns as context (e.g., macroscopic context), and (ii) a spatial self-attention to nearby patches to
disambiguate region-level patterns (e.g., microscopic context).
[19] Since sub-cellular structures can be of different scales in WSis, the spatial encoding models the regional visual patterns that extend beyond the scope of a single patch. The present embodiments incorporate the spatial encoding by a separate self-attention mechanism. Specifically, bidirectional self-attention between all patches within a local region are modeled.
A person having ordinary skill in the art would understand these disclosures to support "the spatial attention models attention to a microscopic visual pattern." Thus, Applicant respectfully submits that "the spatial attention models attention to a microscopic visual pattern"
is supported by the provisional application.
In response “ ("'consideration of the understanding of one skilled in the art in no way relieves the patentee of adequately disclosing sufficient structure in the specification.’ It is not enough for the patentee simply to state or later argue that persons of ordinary skill in the art would know what structures to use to accomplish the claimed function."), quoting Atmel Corp. v. Information Storage Devices, Inc., 198 F.3d 1374, 1380, 53 USPQ2d 1225, 1229 (Fed. Cir. 1999); Biomedino, LLC v. Waters Technologies Corp., 490 F.3d 946, 953, 83 USPQ2d 1118, 1123 (Fed. Cir. 2007) ("The inquiry is whether one of skill in the art would understand the specification itself to disclose a structure, not simply whether that person would be capable of implementing a structure.").” via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(a) Original claims, 7th txt blk.
Upon review of said [14], claim 1’s “models attention” “is not apparent” in [14] and [14] “does not reasonably convey that the inventor had possession of the subject matter”. Since [14] discloses “models…dependencies” and “model…patterns”, wherein “dependencies”39 and “patterns”40 do not mean “attention”41:
[14] In the present embodiments, a transformer-based aggregation model ("WSITrans")
models cross-patch dependencies between all patches to capture microscopic and macroscopic semantic self-attention to the appearance of all other patches in the slide to model slide-level
patterns as context (e.g., macroscopic context), and (ii) a spatial self-attention to nearby patches to
disambiguate region-level patterns (e.g., microscopic context). In addition, an attention-based
confidence regularization is utilized to reduce over-emphasis on single patches for predictions.
The functioning of WSITrans with a tumor-grading classification task and with a survival
prediction regression task are illustrated. These methods outperform existing approaches by at
least 3.59% and 1.64% in ,c-score and C-index, respectively.
Thus, “persons skilled in the art would not recognize in the disclosure a description of the invention defined by the claims”.
Applicants state in page 12:
The Office Action indicates the claimed "the semantic attention models attention to a macroscopic visual pattern" of claim 20 is allegedly not supported in the provisional application. Applicant respectfully disagrees. The provisional application describes:
[14] In the present embodiments, a transformer-based aggregation model ("WSITrans") models cross-patch dependencies between all patches to capture microscopic and macroscopic patterns in WSis. Specifically, WSITrans includes two types of self-attention for each patch: (i) a semantic self-attention to the appearance of all other patches in the slide to model slide-level patterns as context (e.g., macroscopic context), and (ii) a spatial self-attention to nearby patches to disambiguate region-level patterns (e.g., microscopic context).
A person having ordinary skill in the art would understand this disclosure to support "the semantic attention models attention to a macroscopic visual pattern." Thus, Applicant respectfully submits that "the sematic attention models attention to a macroscopic visual pattern"
is supported by the provisional application.
In response “ ("'consideration of the understanding of one skilled in the art in no way relieves the patentee of adequately disclosing sufficient structure in the specification.’ It is not enough for the patentee simply to state or later argue that persons of ordinary skill in the art would know what structures to use to accomplish the claimed function."), quoting Atmel Corp. v. Information Storage Devices, Inc., 198 F.3d 1374, 1380, 53 USPQ2d 1225, 1229 (Fed. Cir. 1999); Biomedino, LLC v. Waters Technologies Corp., 490 F.3d 946, 953, 83 USPQ2d 1118, 1123 (Fed. Cir. 2007) ("The inquiry is whether one of skill in the art would understand the specification itself to disclose a structure, not simply whether that person would be capable of implementing a structure.").” via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(a) Original claims, 7th txt blk.
Upon review of said [14], claim 1’s “models attention” “is not apparent” in [14] and [14] “does not reasonably convey that the inventor had possession42 of the subject matter”. Since [14] discloses “models…dependencies” and “model…patterns”, wherein “dependencies”43 and “patterns”44 do not mean “attention”45:
[14] In the present embodiments, a transformer-based aggregation model ("WSITrans")
models cross-patch dependencies between all patches to capture microscopic and macroscopic semantic self-attention to the appearance of all other patches in the slide to model slide-level
patterns as context (e.g., macroscopic context), and (ii) a spatial self-attention to nearby patches to
disambiguate region-level patterns (e.g., microscopic context). In addition, an attention-based
confidence regularization is utilized to reduce over-emphasis on single patches for predictions.
The functioning of WSITrans with a tumor-grading classification task and with a survival
prediction regression task are illustrated. These methods outperform existing approaches by at
least 3.59% and 1.64% in ,c-score and C-index, respectively.
Thus, “persons skilled in the art would not recognize in the disclosure a description of the invention defined by the claims”.
Applicants state on page 12:
Accordingly, claims 1-20 are entitled to the benefit of the provisional application filed on
October 7, 2021.
In response, “When an explicit limitation [“a global pattern related to the one or more histological features”] in a claim [claim 1] "is not present in the written description [63/253,514] whose benefit is sought it must be shown that a person of ordinary skill would have understood, at the time the patent application was filed, that the description requires that limitation." Hyatt v. Boone, 146 F.3d 1348, 1353, 47 USPQ2d 1128, 1131 (Fed. Cir. 1998)” via:
MPEP2163 Guidelines for the Examination of Patent Applications Under the 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph, "Written Description" Requirement [R-01.2024]
II. METHODOLOGY FOR DETERMINING ADEQUACY OF WRITTEN DESCRIPTION
A. Read and Analyze the Specification for Compliance with 35 U.S.C. 112(a) or Pre-AIA 35 U.S.C. 112, first paragraph
3. Determine Whether There is Sufficient Written Description to Inform a Skilled Artisan That Inventor was in Possession of the Claimed Invention as a Whole at the Time the Application Was Filed
(b) New Claims, Amended Claims, or Claims Asserting Entitlement to the Benefit of an Earlier Priority Date or Filing Date under 35 U.S.C. 119, 120, 365, or 386, 2nd para:
To comply with the written description requirement of 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph, or to be entitled to an earlier priority date or filing date under 35 U.S.C. 119, 120, 365, or 386, each claim limitation must be expressly, implicitly, or inherently supported in the originally filed disclosure. When an explicit limitation in a claim "is not present in the written description whose benefit is sought it must be shown that a person of ordinary skill would have understood, at the time the patent application was filed, that the description requires that limitation." Hyatt v. Boone, 146 F.3d 1348, 1353, 47 USPQ2d 1128, 1131 (Fed. Cir. 1998); see also Akeva LLC v. Nike, Inc., 817 Fed. Appx. 1005, 1012-13, 2020 USPQ2d 10797 (Fed. Cir. 2020) (The court found that the continuation patents were not entitled to the benefit of an earlier filing date because the continuation patents removed a disclaimer that the invention did not cover shoes with conventional fixed rear soles that was present in the earlier filed patents); In re Wright, 866 F.2d 422, 425, 9 USPQ2d 1649, 1651 (Fed. Cir. 1989) (Original specification for method of forming images using photosensitive microcapsules which describes removal of microcapsules from surface and warns that capsules not be disturbed prior to formation of image, unequivocally teaches absence of permanently fixed microcapsules and supports amended language of claims requiring that microcapsules be "not permanently fixed" to underlying surface, and therefore meets description requirement of 35 U.S.C. 112.); In re Robins, 429 F.2d 452, 456-57, 166 USPQ 552, 555 (CCPA 1970) ("[W]here no explicit description of a generic invention is to be found in the specification[,] ... mention of representative compounds may provide an implicit description upon which to base generic claim language."); In re Smith, 458 F.2d 1389, 1395, 173 USPQ 679, 683 (CCPA 1972) (a subgenus is not necessarily implicitly described by a genus encompassing it and a species upon which it reads); Regents of the Univ. of Minnesota v. Gilead Scis., Inc., 61 F.4th 1350, 1356-58, 2023 USPQ2d 269 (Fed. Cir. 2023) ( The court found the later-filed patent claims could not receive benefit under 35 U.S.C. 120, because the earlier-filed applications did not have ipsis verbis disclosure of the claimed subgenus and did not provide sufficient blaze marks to provide the later-filed claims with sufficient support under 35 U.S.C. 112(a).) ;In re Robertson, 169 F.3d 743, 745, 49 USPQ2d 1949, 1950-51 (Fed. Cir. 1999) ("To establish inherency, the extrinsic evidence ‘must make clear that the missing descriptive matter is necessarily present in the thing described in the reference, and that it would be so recognized by persons of ordinary skill. Inherency, however, may not be established by probabilities or possibilities. The mere fact that a certain thing may result from a given set of circumstances is not sufficient.’" (citations omitted)); Yeda Research and Dev. Co. v. Abbott GMBH & Co., 837 F.3d 1341, 120 USPQ2d 1299 (Fed. Cir. 2016) ("Under the doctrine of inherent disclosure, when a specification describes an invention that has certain undisclosed yet inherent properties, that specification serves as adequate written description to support a subsequent patent application that explicitly recites the invention’s inherent properties.") (citing Kennecott Corp. v. Kyocera Int’l, Inc., 835 F.2d 1419, 1423, 5 USPQ2d 1194 (Fed. Cir. 1987)). Furthermore, each claim must include all elements which applicant has described as essential. See, e.g., Johnson Worldwide Assoc. Inc. v. Zebco Corp., 175 F.3d at 993, 50 USPQ2d at 1613; Gentry Gallery, Inc. v. Berkline Corp., 134 F.3d at 1479, 45 USPQ2d at 1503; Tronzo v. Biomet, 156 F.3d at 1159, 47 USPQ2d at 1833.
Since “histological feature” or “models attention” is not recognized in [63/253,514], it is not “ ‘shown that a person of ordinary skill would have understood, at the time the patent application was filed, that the description requires that limitation’ ”.
IV. Claim Rejections – 35 USC 112
Applicant’s arguments, see remarks, page 13, filed 4/27/2026, with respect to 35 USC 112(b) have been fully considered and are persuasive. The 35 USC 112(b) rejection of claims 12,14 has been withdrawn.
V. Claim Rejections – 35 USC 101
1. Claims 1 and 20 are not directed to mathematical concepts
Applicants state in page 14:
Here, amended claim 1 is not directed to a mathematical concept because it does not recite mathematical relationships, mathematical formulas or equations, or mathematical calculations. Although the Office Action points to claim terms such as "embedding," "encoding," "generating," and "representation," and provides dictionary definitions suggesting mathematical connotations (Office Action, pp. 7-8), an embedding as recited in claim 1 is not itself a mathematical concept. Rather, it is a data representation of histological features, used within a computer-implemented process. Claim 1 does not recite any mathematical relationships, formulas, or calculations that define specifically how the embedding is generated or manipulated. The claims recite what is done with the embeddings-incorporating local and global patterns derived from neighboring patches and the WSI as a whole-not mathematical formulas or equations for computing them.
The examiner respectfully disagrees since claim 1’s, interpreted under the broadest reasonable interpretation of applicant’s disclose such as paragraph [104]:
“The scope of this disclosure encompasses all changes, substitutions, variations,
alterations, and modifications to the example embodiments described or illustrated herein that a person having ordinary skill in the art would comprehend. The scope of this disclosure is not limited to the example embodiments described or illustrated herein”
,“embedding” is defined as “the mapping of one set into another”, wherein mapping is defined: Mathematics. function wherein function is defined: a relationship in which an input value of a variable has a specifically calculated output value: for example, if the function of x is x 2 , the output will always be the square of whatever the value of x is. f, F (Dictionary.com). Thus claim 1 recites a mathematical concept under the broadest reasonable interpretation of claim 1.
Applicant’s state in page 15:
According to Example 39, "the claim does not recite any mathematical relationships, formulas, or calculations. While some of the limitations may be based on mathematical concepts, the mathematical concepts are not recited in the claims." (Subject Matter Eligibility Examples 37 to 42, page 9.) As further explained in the August memorandum, "[e]ven though "training the neural network" involves a broad array of techniques and/or activities that may involve or rely upon mathematical concepts, the limitation does not set forth or describe any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols." (August Memorandum, "Reminders on Evaluating Subject Matter Eligibility of Claims Under 35 U.S.C. § 101," page 3.)
Here, claim 1 is analogous to USPTO Eligibility Example 39. Claim 1 recites limitations such as "generating, by a neural network model, an encoded embedding for the patch by: incorporating into the extracted embedding a local pattern related to the one or more histological features, wherein the local pattern is derived from embeddings of one or more neighboring patches; and incorporating into the extracted embedding a global pattern related to the one or more histological features, wherein the global pattern is derived from embeddings of the WSI as a whole." These limitations do not set forth or describe any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols. Accordingly, consistent with MPEP § 2106.04(a)(2), the August Memorandum, and USPTO Example 39, claim 1 does not recite a mathematical concept. Claim 20 recites limitations similar to claim 1, and thus does not recite a mathematical concept for at least the same reasons as claim 1.
The examiner respectfully disagrees since example 39 does not have a no-limit- in-scope46 statement as present in applicant’s disclosure at said paragraph [104]:
“The scope of this disclosure encompasses all changes, substitutions, variations,
alterations, and modifications to the example embodiments described or illustrated herein that a person having ordinary skill in the art would comprehend. The scope of this disclosure is not limited to the example embodiments described or illustrated herein”.
Subject Matter Eligibility Examples: Abstract Ideas
Example 39 - Method for Training a Neural Network for Facial Detection
Background:
Facial detection is a computer technology for identifying human faces in digital images. This technology has several different potential uses, ranging from tagging pictures in social networking sites to security access control. Some prior methods use neural networks to perform facial detection. A neural network is a framework of machine learning algorithms that work together to classify inputs based on a previous training process. In facial detection, a neural network classifies images as either containing a human face or not, based upon the model being previously trained on a set of facial and non-facial images. However, these prior methods suffer from the inability to robustly detect human faces in images where there are shifts, distortions, and variations in scale and rotation of the face pattern in the image.
Applicant’s invention addresses this issue by using a combination of features to more robustly detect human faces. The first feature is the use of an expanded training set of facial images to train the neural network. This expanded training set is developed by applying mathematical transformation47 functions48 on an acquired set of facial images. These transformations49 can include50 affine transformations51, for example, rotating52, shifting, or mirroring or filtering transformations, for example, smoothing or contrast reduction. The neural networks are then trained with this expanded training set using stochastic learning with backpropagation which is a type of machine learning algorithm that uses the gradient of a mathematical loss function to adjust the weights of the network. Unfortunately, the introduction of an expanded training set increases false positives when classifying non-facial images. Accordingly, the second feature of applicant’s invention is the minimization of these false positives by performing an iterative training algorithm, in which the system is retrained with an updated training set containing the false positives produced after face detection has been performed on a set of non-facial images. This combination of features provides a robust face detection model that can detect faces in distorted images while limiting the number of false positives.
Claim:
A computer-implemented method of training a neural network for facial detection comprising:
collecting a set of digital facial images from a database;
applying one or more transformations5354 to each digital facial image including55 mirroring, rotating56, smoothing, or contrast reduction to create a modified set of digital facial images;
creating a first training set comprising the collected set of digital facial images, the modified set of digital facial images, and a set of digital non-facial images;
training the neural network in a first stage using the first training set;
creating a second training set for a second stage of training comprising the first
training set and digital non-facial images that are incorrectly detected as facial images after the first stage of training; and
training the neural network in a second stage using the second training set.
2. Claims 1 and 20 are not directed to a mental process.
Applicants state in pages 16,17:
These limitations cannot be performed in the human mind. The human mind is not equipped to extract embeddings representing histological features from patches of a whole slide image, nor to generate encoded embeddings by incorporating local patterns derived from embeddings of neighboring patches and global patterns derived from embeddings of the WSI as a whole. These operations cannot be performed by a human mind because a human mind cannot execute or operate a neural network model. As the August Memorandum states, "[c]laim limitations that encompass AI in a way that cannot be practically performed in the human mind do not fall within this grouping." Claim 1 recites such limitations. Thus, claim 1 is not directed to a mental process. Claim 20 recites limitations similar to claim 1, and thus does not recite a mental process for at least the same reasons as claim 1.
The examiner respectfully disagrees since the claimed “,57 neural network model,” “does not limit the scope of a claim under the broadest reasonable claim interpretation”58 of claim 1.
3. Claims 1 and 20 are not directed to a method of organizing human activity.
Applicants state in page 18, 2nd para:
Claim 1 does not regulate or manage human behavior, does not involve social activities or interactions between people, and does not recite rules or instructions for humans to follow. Instead, claim 1 is directed to a computer-implemented method for analyzing a whole slide image using a neural network model that generates encoded embeddings by incorporating local and global patterns derived from embeddings. This falls outside the enumerated sub-groupings of "certain methods of organizing human activity." Accordingly, claim 1 is not directed to a method of organizing human activity under MPEP § 2106.04(a)(2). Claim 20 recites limitations similar to claim 1, and thus does not recite a method of organizing human activity for at least the same
reasons as claim 1.
The examiner respectfully disagrees since claim 1 claims “performing a pathological task5960” which comprises constitutional rules.
4. Claims 1 and 20 integrate any alleged abstract idea into a practical application
Applicants state in page 21, 2nd para:
Thus, the specification identifies the technical problem of prior art models treating each patch independently without modeling biological context is solved by the specific technical solution of generating encoded embeddings that incorporate both local patterns from neighboring patches and global patterns from the WSI as a whole. This is a structural improvement to how the aggregation model itself functions. Claim 1 reflects this improvement by reciting "for each of the patches, generating, by a neural network model, an encoded embedding for the patch by: incorporating into the extracted embedding a local pattern related to the one or more histological features, wherein the local pattern is derived from embeddings of one or more neighboring patches; and incorporating into the extracted embedding a global pattern related to the one or more histological features, wherein the global pattern is derived from embeddings of the WSI as a whole."
The examiner respectfully disagrees since the claimed “,61 neural network model,” “does not limit the scope of a claim under the broadest reasonable claim interpretation”62 of claim 1.
Applicants state in pages 21,22:
Accordingly, even assuming arguendo that claim 1 is directed to a judicial exception (which Applicant does not concede), claim 1 integrates the alleged judicial exception into a practical application because it recites elements that reflect an improvement to a technology or technical field, and specifically, an improvement to how multiple-instance learning models aggregate patch-level information by encoding each embedding with both local and global contextual patterns, thereby capturing biological context that prior art models failed to model.
The examiner respectfully disagrees since the claimed “,63 neural network64 model,65” “does not limit the scope of a claim under the broadest reasonable claim interpretation”66 of claim 1.
In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “by a first encoder” (no surrounding commas unlike claim 20’s -- ,67 by a first encoder ,68 --), applicant’s remarks, page 22:
Claim 20 recites similar limitations with the additional specificity that the local pattern is incorporated by a first encoder attending to embeddings of nearby patches, and the global pattern is incorporated by a second encoder, further defining the architecture of the neural network model. This is analogous to claims found eligible in Ex parte Carmody (Appeal Docket No. 2025-002843), where the PTAB held that specifying a modular, multi-model architecture enabled improvements over conventional approaches and rendered the claims patent-eligible. In Carmody, the Board found that claims reciting "a plurality of modular plug-and-play tactic-specific models" were patent-eligible because the specification explained, "this modular approach to tactic recommendation (i.e., with a separate model for each tactic) enables the model for each tactic to be updated and improved separately and independently from other tactic-specific models, and also enables models for new tactics to be easily incorporated." Claim 20 is particularly analogous to Carmody because it specifies a multi-encoder architecture within the neural network model. A comparison between claim 20 and Carmody is below:
) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “by a second encoder” (no surrounding commas unlike claim 20’s -- ,69 by a second encoder ,70 --), applicant’s remarks, page 22:
Claim 20 recites similar limitations with the additional specificity that the local pattern is incorporated by a first encoder attending to embeddings of nearby patches, and the global pattern is incorporated by a second encoder, further defining the architecture of the neural network model. This is analogous to claims found eligible in Ex parte Carmody (Appeal Docket No. 2025-002843), where the PTAB held that specifying a modular, multi-model architecture enabled improvements over conventional approaches and rendered the claims patent-eligible. In Carmody, the Board found that claims reciting "a plurality of modular plug-and-play tactic-specific models" were patent-eligible because the specification explained, "this modular approach to tactic recommendation (i.e., with a separate model for each tactic) enables the model for each tactic to be updated and improved separately and independently from other tactic-specific models, and also enables models for new tactics to be easily incorporated." Claim 20 is particularly analogous to Carmody because it specifies a multi-encoder architecture within the neural network model. A comparison between claim 20 and Carmody is below:
) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “the neural network” (no surrounding commas unlike claim 20’s -- ,71 by a neural network model,72 --), applicant’s remarks, page 22:
Claim 20 recites similar limitations with the additional specificity that the local pattern is incorporated by a first encoder attending to embeddings of nearby patches, and the global pattern is incorporated by a second encoder, further defining the architecture of the neural network model. This is analogous to claims found eligible in Ex parte Carmody (Appeal Docket No. 2025-002843), where the PTAB held that specifying a modular, multi-model architecture enabled improvements over conventional approaches and rendered the claims patent-eligible. In Carmody, the Board found that claims reciting "a plurality of modular plug-and-play tactic-specific models" were patent-eligible because the specification explained, "this modular approach to tactic recommendation (i.e., with a separate model for each tactic) enables the model for each tactic to be updated and improved separately and independently from other tactic-specific models, and also enables models for new tactics to be easily incorporated." Claim 20 is particularly analogous to Carmody because it specifies a multi-encoder architecture within the neural network model. A comparison between claim 20 and Carmody is below:
) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “machine learning” in applicant’s remarks, page 25:
As in Carmody73, claim 20 does not merely invoke machine learning74 at a high level. By reciting "incorporating, by a first encoder, a microscopic visual pattern related to the one or more histological features into the extracted embedding corresponding to the patch by attending to the embeddings of nearby patches in the set" and "incorporating, by a second encoder, a macroscopic visual pattern over the WSI as a whole into the extracted embedding corresponding to the patch," claim 20 recites a specific multi-encoder architecture in which separate encoders are each responsible for incorporating two types of contextual information, local patterns and global patterns, respectively. This is analogous to the modular approach found eligible in Carmody, where separate models each handled a distinct function. Accordingly, claim 20 integrates any alleged judicial exception into a practical application by claiming a particular architectural solution to the technical problem of modeling biological context during WSI analysis.
) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
VI. Claim Rejections - 35 U.S.C. § 103
Applicant’s arguments, see remarks, pages 27,28,29 filed 4/27/2026, with respect to the rejection(s) of claim(s) 1,20 including 18,19 under 35 USC 103 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of:
Claim(s) 1,15,16,17 and 18 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1); and
Claim(s) 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17,18,19 further in view of Chiu et al. (US 2020/0357143 A1) further in view of LIU et al. (CN 112163608 A) with SEARCH machine translation, wherein Wainrib teaches “neighboring tiles” [0079] 1st S as claim 1’s “neighboring patches” as indicated in fig. 6’s 620: WEIGHTED POOLING & 614: CONCAT FEATURES:
[0079] In an alternative embodiment, the device can also aggregate clusters of neighboring tiles. In this embodiment, aggregating a cluster of tiles can include concatenating the tiles of the cluster, selecting a single tile from the cluster according to a given criterion, using the cluster as a multidimensional object, or aggregating the values for example through a mean or a max pooling operation. In addition, the device can apply an autoencoder on the extracted feature vectors so as to reduce the dimensionality of the features. In one embodiment, the image can be a histopathology slide, the region of interest being a tissue region, and the classification of the image being a diagnosis classification.
PNG
media_image5.png
700
1009
media_image5.png
Greyscale
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically75 disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1,15,16,17 and 18 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1).
PNG
media_image6.png
899
483
media_image6.png
Greyscale
Re 1. (Currently Amended) SALTZ teaches A computer-implemented method for analyzing a whole slide image (WSI) (or “representative H&E diagnostic whole-slide images (WSIs)” [0033]: fig. 1C: 57, a patch 57 with smaller patches 51, is of the WSIs in fig. 1B:38) in light of biological context, (likewise) comprising76:
extracting77 (via “a Convolutional Autoencoder (CAE) for fully unsupervised, simultaneous nucleus detection and feature extraction in histopathology tissue images” [0080]: fig. 1C:72: “Sparse Auto Encoder”: fig. 2A:72: “Sparse Autoencoder”: fig. 5:151: “Train an unsupervised Convolutional Autoencoder (CAE)”) an embedding78 (or likewise fig. 8G’s mapping from data-points to an image at “Step 5” & “an identity mapping layer” [0379]: fig. 2B:“Part 6”) for (“the center of”) each (“patch” [0098] 1st S: fig. 1C:51:squares) of a set (or “four datasets” [0081] penult S) of patches sampled (via “tumoral tissue samples” [0079] last S) from a WSI (or “computationally stained and digitized whole slide images of Hematoxylin and Eosin (H&E) stained pathology specimens obtained from biopsied tissue” [0079] penult S), wherein the embedding (mapping-layer) represents one or more histological features (or “histopathology tissue image” “feature” [0080] 3rd S) of the respective (center) patch (51) of the WSI;
for each of the patches (51) of the set of patches, generating 7980 (“encoded”81 is an operation (relying/retransmit a message, Morse code, to the brain as if by means of a telegraphic relay) of the human brain and understood given that SALTZ teaches “a convolutional neural network82 (CNN)” that models/simulates/mimics the operation (encode) of human brain) embedding (or likewise “encode not only appearance, but also spatial information in feature maps” [0134]) for the patch by:
incorporating (or classifying forming a group as shown in fig. 14A by tumor type, such as “Non-Brisk Focal” tumor type) into the extracted embedding83 (also forming said classification group) nd S: fig. 12D:304: “Non-Brisk Focal” category or the tumor type representing the local patterns: two local patterns are shown in fig. 12D:304 and further classified or incorporated in said group in fig. 14A as the upper-right dark rectangle at classification coordinates (“BRCA”, “Non-Brisk Focal”)) related to the one or more histological features, wherein the local pattern is derived from embeddings of one or more neighboring patches84; and
incorporating into the extracted embedding (the claimed incorporating and embedding have the same result of classifying into any of the rectangles in fig. 14A) tumor”, [0339] 2nd S) related to the one or more histological features, wherein the global pattern is derived from embeddings of
generating a representation (via “generate the foreground reconstructed image85 87 and background reconstructed image 89” [0145] penult S) for the WSI by combining (via “Finally the two intermediate images 87, 89 are summed to form the final reconstructed image at step 90” [0145] last S) the encoded (via said CAE)
performing a pathological (“pipeline” [0084] 1st S) task (via a “pathology” “classification model” [0079] 2nd S: fig. 1A:5: “Trained Model”) based on the representation for the WSI.
SALTZ does not teach the difference86 of claim 1 of:
neighboring (patches87)88.
PNG
media_image7.png
693
1062
media_image7.png
Greyscale
PNG
media_image8.png
354
778
media_image8.png
Greyscale
Wainrib teach the difference89 of claim 1 of:
neighboring (patches)90 (or likewise “neighboring tiles” [0079] 1st S).
Since SATLZ teaches not having enough “memory”, [0134],2nd S:
[0134] In order to perform unsupervised representation learning and detection, the system modifies the conventional CAE to encode not only appearance, but also spatial information in feature maps. In this regard, the CAE first learns to separate background 74 (e.g. cytoplasm) and foreground 73 (e.g. nuclei 76) in an image patch 71, as also shown hereinbelow in FIG. 2B. It is noted that an image patch is a rectangular region in a whole slide tissue image in certain embodiments or aspects. The CAE 70 implements image patches, because a tissue image can be very large and may not fit in memory. Hence, working with and processing image patches 71 is computationally more efficient. It is common in tissue image analysis to partition tissue images into patches and process the patches. Hence, the partitioned image patches may also be referred herein simply as the images. The CAE encodes the input image in a set of low resolution feature maps (background feature maps 74) with a small number of encoding neurons. The feature maps generally encode large scale color and texture variations because of their limited capacity and to increase computational efficiencies. Thus, these feature maps 74 encode the image background. The high frequency residual between the input image 71, 81 and the reconstructed background 89 during the sparse autoencoding step 72, is the foreground that contains nuclei 76, hence, the reconstructed image 75 is output.
one of skill in the art would have looked to other teachings of memory as a solution to the memory problem -- such as Wainrib’s “it is difficult to store…inside the random-access memory of a computer” [0094], 5th S:
[0094] Process 200 tiles the image into a set of image tiles at block 206. In one embodiment, process 200 uses the tiling increase the ability of preprocessing the images. For example and in one embodiment, using a tiling method is helpful in histopathology analysis, due to the large size of the whole-slide image. More broadly, when working with specialized images, such as histopathology slides, the resolution of the image sensor used in these fields can grow as quickly as the capacity of random-access memory associated with the sensor. With this increased image size, it is difficult to store batches of images, or sometimes even a single image, inside the random-access memory of a computer. This difficulty is compounded if trying to store these large images in specialized memory of a Graphics Processing Unit (GPU). This situation makes it computationally intractable to process an image slide, or any other image of similar size, in its entirety.--
and thus make SALTZ’s be as Wainrib’s seeing in the change “a reduction of resources further improves the performance of the device when executing the image classification task. In addition, the device can classify a whole-slide image, even when this type of image is too large to fit in the memory of a graphics processing unit commonly used to train machine learning models. In a further embodiment, the device reduces the dimensionality of the data, thus giving better generalization error and is more efficient in terms of model accuracy.”, Wainrib [0074] last Ss, via explicit creative or even routine steps via:
a) create a tile extraction program based on Wainrib’s fig. 2:
PNG
media_image9.png
1205
862
media_image9.png
Greyscale
b) create a thresholding program based on Saltz’s fig. 1D;
b1) write code at fig. 1D, step 61 calling Wainrib’s tile extraction program of fig. 2 executing steps 202-210 and then return to Saltz’s thresholding program of fig. 1D to continue with ten scored-tiles/patches to step 62: “Label the patches”:
PNG
media_image10.png
985
1033
media_image10.png
Greyscale
Re 15. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib teaches The method of claim 1, wherein combining the encoded
Re 16. Currently Amended), SALTZ of the combination of SALTZ, Wainrib teaches The method of claim 1, wherein performing a pathological task based on the representation for the WSI comprises:
classifying the one or more histological features extracted from the WSI;
classifying a pathological type (“in clinical analysis and/or prognosis and result in more accurate patient summaries including more accurate classification of the respective cancer type” [0205] 4th S) of the WSI;
predicting a progression risk of a disease associated with the one or more histological features; or
determining a diagnosis (via “ extract, quantify, characterize and correlate TIL Maps using digitized H&E stained diagnostic tissue slides that are routinely obtained as part of cancer diagnosis”, SALTZ [0016]) of a patient associated with the WSI.
Re 17. (Original), SALTZ of the combination (illustrated above) of SALTZ, Wainrib teaches The method of claim 16, wherein the pathological task may be performed using a classifier model (“in which deep classification learning models (for example, lymphocyte infiltration classification CNN and a necrosis segmentation algorithm) are implemented to generate tumor infiltrating lymphocyte maps that are useful in generating prognostic values in diagnosis and/or related classification”, SALTZ [0019] last S) or a regressor model.
Claim 18 is rejected like claim 1:
Re 18. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib teaches One or more computer-readable non-transitory91 storage media (SALTZ: fig. 16:320: “Machine-readable medium(s)”) embodying software for analyzing a whole slide image (WSI) in light of biological context, the software comprising instructions operable when executed to:
extract an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches, generate, by a neural network, an encoded embedding for the patch by:
incorporating into the extracted embedding wherein the local pattern is derived from embeddings of one or more neighboring patches; and
incorporating into the extracted embedding related to the one or more histological features, wherein the global pattern is derived from embeddings of
generate a representation for the WSI by combining the encoded
perform a pathological task based on the representation for the WSI.
Claim 19 is rejected like claims 1 and 18:
Re 19. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib teaches A system for analyzing a whole slide image (WSI) in light of biological context comprising one or more processors (SALTZ: fig. 16: 304: “Processing Device(s)”) and a memory coupled to the processors comprising instructions executable by the processors, the processors being operable when executing the instructions to:
extract an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches of the set of patches, generate, by a neural network model, an encoded embedding for the patch by:
incorporating into the extracted embedding wherein the local pattern is derived from embeddings of one or more neighboring patches; and
incorporating into the extracted 92 related to the one or more histological features, wherein the global pattern is derived from embeddings of
generate a representation for the WSI by combining the encoded
perform a pathological task based on the representation for the WSI.
Claim(s) 2 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Tu et al. (US 2023/0070286 A1):
PNG
media_image11.png
899
483
media_image11.png
Greyscale
PNG
media_image12.png
1464
1138
media_image12.png
Greyscale
Re 2. (Original), SALTZ of the combination (illustrated above) of SALTZ, Wainrib teaches The method of claim 1, wherein the patches are sampled (resulting in “ten patches sampled from each slide” [0166] last S: fig. 3A: step B) by applying a hierarchical sampling strategy (“in step 182” [0269] 3rd S: fig. 5A) to a randomly selected plurality (“8 slides” [0220]) of clusters of the patches.
SALTZ of the combination (illustrated above) of SALTZ, CASALE does not teach the difference of claim 2 of “hierarchical” (sampling strategy).
Tu teaches the difference of claim 2 of:
hierarchical (“sampling iterations” [0019] last S) (sampling strategy).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib teaches sampling, one of skill in the art of sampling can make SALTZ’s of the combination of SALTZ, Wainrib be as Tu’s seeing in the change “hierarchical sampling to iteratively refine clusters and better preserve shape details and structural relationships” Tu, [0037] penult S.
Claim(s) 3,4 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Tu et al. (US 2023/0070286 A1) as applied in claim 2 further in view of Martinez Manzano et al. (US 2023/0401590 A1):
PNG
media_image13.png
899
697
media_image13.png
Greyscale
Re 3. (Original), SALTZ of the combination of SALTZ, Wainrib,Tu teaches The method of claim 2, further comprising applying the hierarchical sampling strategy by:
for each of the randomly selected (“8 slides” Saltz: [0220]: fig. 3D:556: “Randomly select 8 slide images from each group (A-G)” & “examples 135 of unsupervised nucleus detection 137 and foreground 138, background image representation 139 and reconstruction results 140 of using crosswise sparse CAEs” [0252] 1st S: fig. 4A) clusters93 (“of TIL patches derived from the affinity propagation clustering of the TIL patches” [0312] penult S):
randomly sampling (resulting in “randomly sampled patches” [0230] 4th S) a centroid of the (TIL: Tumor-Infiltrating Lymphocytes) cluster (comprising a “cluster” “central representative” [0312] last S: fig. 12A-D:right cluster column);
for each of the (TIL) patches in the (central representative) cluster, determining a distance of the patch to (i.e., “variance” “dispersion” “In terms of TIL patch distances to a given cluster center” [0320] 3rd S) the centroid; and
randomly sampling (resulting in “randomly sampled patches” [0230] 4th S) all patches (resulting in “randomly sampled patches” [0230] 4th S) in the (TIL) cluster having a (center-patch-distance-variance-dispersion) distance to the centroid within a threshold distance.
SALTZ of the combination of SALTZ, Wainrib,Tu does not teach the difference of claim 3 of:
clusters94…
a centroid…
the centroid…
the centroid within a threshold distance.
Martinez teaches the difference of claim 3 of:
(“randomly select a number of” [0033] 6th S) clusters95 (fig. 5:502,506,510,514,518: smaller circles)…
a centroid (“data point for each of the groups generated by the clustering technique” [0033] penult S: fig. 5:540,508,512,516,520)…
the centroid (“will be calculated” [0033] last S)…
the centroid (clusters) within a threshold distance (“of the…closest cluster” [0155] 2nd S: fig. 5:526,528,530,532: distances).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Tu teaches a cluster, one of skill in the art of clusters can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Tu be as Martinez’s seeing the change resulting in reliable and “stabilized”96 “clusters”, Martinez [0094] 3rd S.
Re 4. (Original), SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Tu, Martinez teaches The method of claim 3, wherein the threshold distance is based on (via th combination (illustrated above) of SALTZ, Wainrib,Tu, Martinez) the pathological task.
Claim(s) 5,9 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1):
PNG
media_image14.png
899
697
media_image14.png
Greyscale
Re 5. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib teaches The method of claim 1, wherein incorporating into the extracted embedding the local pattern comprises using a first (or likewise fig. 1C:58: “Necrosis CNN on 6.67x patch”) encoder (via a “spatial information” “CAE” [0134]: Spatial-info Convolutional Auto Encoder) to encode (via Morse code axon-destination nerve impulse telegraphing neurons as indicated in SALTZ’s fig. 2C) the extracted embedding by attending to embeddings of one [[or more]] nearby (“local population” SALTZ [0171] last S) patch[[es]] (said 6.67x patch) in the set (or “four datasets”, SALTZ [0081] penult S).
SALTZ of the combination of SALTZ, Wainrib does not teach the difference of claim 5 of:
by attending to (embeddings)97.
Chiu teaches the difference of claim 5:
by attending to (“the transformed maps from Equation (1), above” [0042]) (embeddings)98.
Since SALTZ of the combination of SALTZ, Wainrib teaches feature extraction, one of skill in the art of feature extraction can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib be as Chiu’s seeing the change “improved visual localization”, Chiu [0010] 1st S:
PNG
media_image15.png
2109
1137
media_image15.png
Greyscale
Claim 9 is rejected like claim 5:
Re 9. Currently Amended), SALTZ of the combination of SALTZ, Wainrib,Chiu teaches The method of claim 1, wherein incorporating into the extracted embedding a global pattern comprises using a second (or likewise fig. 1C:53: “Lymphocyte CNN on 20x patches”) encoder (via said CAE) to encode the extracted embedding by attending to embeddings of all other patches (or likewise “each99 patch” SALTZ [0098]) in the set (of “four datasets” SALTZ [0081] penult S).
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Xiao et al. (US 2023/0274248 A1):
PNG
media_image16.png
899
697
media_image16.png
Greyscale
Re 6. (Original), SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches The method of claim 5, wherein the one or more nearby (“local population” [0171]) patches are defined (as shown by the patch-outlines in fig. 1C:51) as those within a maximum relative distance corresponding to a specified pathological type (of “13 cancer types” [0086] 2nd S) of the WSI.
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu does not teach the difference of claim 6 of:
(patches are defined) as… within a maximum relative distance corresponding to (a specified pathological type).
Xiao teaches the difference of claim 6 of:
(patches are defined) as… within a (”local” [0041] 1st S) maximum (fig. 8:804b,8o4c) relative distance (fig. 8:810) corresponding to (“an item” [0041] 1st S: fig. 8:302) (a specified pathological type).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches counting, one of skill in the art of counting can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Chiu be as Xiao’s seeing the change “accurate, real-time counts”, Xiao [0018] penult S.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Bassi (US 2006/0050074 A1):
PNG
media_image17.png
899
697
media_image17.png
Greyscale
Re 7. (Currently Amended), SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches The method of claim 5, wherein input for the first encoder comprises:
a (“center” [0098] 1st S) position of the corresponding patch (fig. 1C:51,57) and a sequence (fig. 4A: “1st”, [0241] 1st S, to 5th images) of absolute (“center” [0098] 1st S) positions of the nearby (“local population” [0171]) patches.
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu does not teach the difference of claim 7 of “absolute”.
Bassi teaches the difference of claim 7 of absolute (“position of patch origin”, Bassi [0056]).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches a patch, one of skill in the art of patches can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Chiu be as Bassi’s seeing in the change “correcting small distortions in projectors, cameras, and display devices, to correcting for perspectives like keystone or special wide-angle lens corrections, and to a complete change in image geometry such as forming rectangular panoramas from circular 360 degree images, or other rectangular to polar type mappings.”, Bassi [0028] last S, such as corrections to the lens of microscopes and “optical see-through display”, SALTZ [0391] 2nd S, thereof.
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of Bassi (US 2006/0050074 A1) as applied in claim 7 further in view of OGASAWARA et al. (US 2023/0316489 A1):
PNG
media_image18.png
899
905
media_image18.png
Greyscale
Re 8. (Original), SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu, Bassi teaches The method of claim 7, wherein the absolute positions (“position of patch origin”, Bassi [0056]) are normalized to correspond to a standard level100 of (“20x”) magnification (“level” SALTZ [0107] 1st S).
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu, Bassi does not teach the difference of claim 8 of:
(the absolute positions) are normalized to correspond to a standard level101 of (magnification).
OGASAWARA teaches the difference of claim 8 of:
(the absolute positions) (“images” [0094], annotated below) are normalized to correspond to a standard (“magnification ratio” [0094] annotated below) level102 of (magnification).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu, Bassi teaches magnification, one of skill in the art of magnification can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib ,Chiu,Bassi be as OGASAWARA’s seeing in the change “variations in image resolution…suppressed, and discrimination accuracy…improved”, OGASAWARA [0094], below:
PNG
media_image19.png
1801
871
media_image19.png
Greyscale
Claim(s) 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied to claims 5,9 further in view of Hossain et al. (Bi-SAN-CAP: Bi-Directional Self-Attention for Image Captioning):
PNG
media_image20.png
899
905
media_image20.png
Greyscale
Re 10. (Currently Amended), SALTZ of the combination (illustrated above) of SALTZ, CASALE, Chiu teaches The method of claim 9, wherein the second encoder (via said CAE) is a bidirectional self-attention (CAE) encoder with multi-head attention layers
SALTZ of the combination (illustrated above) of SALTZ, Wainrib, Chiu does not teach the difference of claim 10 of:
a bidirectional self-attention (encoder) with multi-head attention (layers).
Hossain teaches the difference of claim 10 of:
a (“modified”, 3rd pg.: III Model Architecture, 1st para, 2nd S: fig. 2) bidirectional self-attention (encoder) with multi-head attention (“Multi-Head Attention” 4th p) (layers).
Since SALTZ of the combination of SALTZ, Wainrib,Chiu teaches a query, SALTZ:
[0224] It is noted that each patch in a WSI is represented as a rectangle and associated with a classification label and the probability value computed by the CNN. This information is stored as a data element (document) in FeatureDB and indexed to speed up queries by the TIL-Map editor to retrieve and display subsets of patches. After classification results for a set of WSIs have been loaded to the database, a pathologist can use a web browser to view and update the classification results. The pathologist may implement the TIL-Map editor to examine an image, query FeatureDB to retrieve patches visible within the view point and zoom level and display them as a two-color heatmap. The pathologist can edit the heatmap using the “Lymphocyte Sensitivity,” “Necrosis Specificity,” “Smoothness” sliders in a panel 255 (for example, as shown in FIGS. 7E-7F). These slides 255 permit the pathologist to change the threshold value which determines if a patch should be classified as lymphocyte-infiltrated or not.
, one of skill in the art of queries can make SALTZ’s of the combination of SALTZ, Wainrib,Chiu be as Hossain’s seeing the change being “useful to…search…queries”, Hossain, 1st page, I. Introduction, 1st para, 2nd S.
Claim(s) 11,12 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied to claims 5,9 further in view of LI et al. (CN 113269724 A) with SEARCH machine translation:
PNG
media_image21.png
899
905
media_image21.png
Greyscale
Re 11. (Currently Amended), SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches The method of claim 9, wherein input for the second (or likewise fig. 1C:53: “Lymphocyte CNN on 20x patches”) encoder comprises the embeddings of the other patches in the set and a learnable token.
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu does not teach the difference of claim 11 of:
a learnable token.
LI teaches the difference of claim 11:
a learnable token (or likewise “adding a cls-token… so as to …input to the Transfomer103”, pg. 4, 4th txt blk, “to learn context and produce output”).
Since SALTZ of the combination of SALTZ, Wainrib,Chiu teaches “state-of-the-art nucleus analysis methods are” known such as “CNN… classification”: a classifier, SALTZ:
[0083] Collecting a large-scale supervised dataset is a labor intensive and challenging process since it requires the involvement of expert pathologists whose time is a very limited and expensive resource. Thus, in accordance with an embodiment, existing state-of-the-art nucleus analysis methods are semi-supervised, so disclosed is an unsupervised representation method that employs the following steps: 1). Pre-train an autoencoder for unsupervised representation learning; 2). Construct a CNN from the pretrained autoencoder; and 3). Fine-tune the constructed CNN for supervised nucleus classification. In order to better capture the visual variance of nuclei, one usually trains the unsupervised autoencoder on image patches with nuclei in the center. This requires a separate nucleus detection step which in most cases, requires further fine-tuning to optimize the final classification performance.
, one of skill in the art of classifiers “would have...or could have done”104 is look to other known teachings regarding CNN classification nucleus analysis and thus can make SALTZ’s of the combination of SALTZ, Wainrib,Chiu be as Li’s seeing the change, Li, page 4 ,last txt blk continued to page 5:
Compared with the prior art, the present invention has at least the following beneficial technical effects:
1, auxiliary pathological doctor for classifying the cancer subtype, improving the working efficiency of the doctor.
2, learning based on example, more attention cell and its surrounding environment characteristic, ignoring the meaningless background area, reducing the calculation complexity, improving the classification accuracy.
3, based on the model, it can better capture the pathological characteristics of the fine particle size, such as cell level and cell level characteristic, more accurately performing the subtype classification.
via explicit creative or even routine steps:
a) create a tile extraction program based on Wainrib’s fig. 2:
PNG
media_image9.png
1205
862
media_image9.png
Greyscale
b) create a thresholding program based on Saltz’s fig. 1D;
b1) write code at fig. 1D, 10 patch extraction step 61 calling Wainrib’s tile extraction program of fig. 2 executing steps 202-210 and then return to Saltz’s thresholding program of fig. 1D to continue with ten scored-tiles/patches to step 62: “Label the patches”:
b2) write code at fig. 1D:60: “Apply trained CNN on all slides” calling LI’s transformer program of fig. 1 (step “c)” below)
PNG
media_image10.png
985
1033
media_image10.png
Greyscale
c) create a computer transformer program based on LI’s fig. 1:
c1) write code inputting SALTZ’s lymphocyte CNN (fig. 1C: “Lymphocyte CNN on 20x patches” into LI’s fig. 1’s “Small convolutional neural network”:
PNG
media_image22.png
1400
1115
media_image22.png
Greyscale
c2) return to SALTZ’s thresholding program of fig. 1D, step 60 using the transformer results of Li’s fig. 1 for the 10 patch extraction step 61 of SALTZ’s thresholding program of fig. 1D;
d) run all programs inputting images as required to the programs.
Re 12. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib,Chiu,Li teaches The method of claim 11, wherein, during a training phase (i.e., “learning”, Li, page 4, last txt blk, bullet “2”), generating a representation (or “spatially characterizing TIL Maps”, SALTZ [0079] penult S) of the WSI based on the encoded based on the learnable token (or likewise “adding a cls-token…so as to generate a sequence SO, which specifically represents the…image-level classification characteristic”, LI, page 6, last txt blk.
Claim(s) 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of KAZUHIRO (JP 2016-158059 A) with SEARCH machine translation:
PNG
media_image23.png
899
905
media_image23.png
Greyscale
Re 13. (Currently Amended), SALTZ of the combination of SALTZ, Wainrib,Chiu teaches The method of claim 9, further comprising:
the to the embedding of all the other patches in the set to reduce overemphasis on a few (“TCGA tumor types” (The Cancer Genome Atlas) SALTZ [0090]) of the patches (via SALTZ: fig. 1B: 34: “Extract patches from marked regions”) to generate the representation (via “generate the foreground reconstructed image105 87 and background reconstructed image 89” SALTZ [0145] penult S) for the WSI.
SALTZ of the combination of SALTZ, Wainrib ,Chiu does not teach the difference of claim 13 of:
106…107 to reduce overemphasis on (a few).
KAZUHIRO teaches the difference of claim 13 of:
108…109 to reduce overemphasis (“according to the overemphasis suppression”, pg. 16, penult txt blk) on (a few).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches an image, one of skill in the art of images can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Chiu be as KAZUHIRO’s seeing in the change “image quality…improved”, KAZUHIRO, pg. 18, 1st txt blk.
Claim(s) 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17 and 18 and 19 further in view of Chiu et al. (US 2020/0357143 A1) as applied in claims 5,9 further in view of KAZUHIRO (JP 2016-158059 A) with SEARCH machine translation as applied in claim 13 further in view of Jaiswal et al. (US 11,775,617 B1):
PNG
media_image24.png
899
905
media_image24.png
Greyscale
Re 14. (Currently Amended), SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu,KAZUHIRO teaches The method of claim 13, wherein regularizing (via “normalized”-“embeddings” Chiu [0045] last S: fig. 1:150) the attention to the embedding of all the other patches in the set comprises:
calculating an attention map (comprised by “The spatial attention maps” Chiu [0042] last S) STAGE 1“: “GENERATING EMBEDDINGS FROM IMAGE DATA”) corresponding to patches sampled from the WSI; and
adding a negative entropy of the attention map (comprised by “The spatial attention maps” Chiu [0042] last S) to a training objective (“to train the semantic embedding space”, Chiu [0047] 1st S, as an aimed training goal: fig. 1:140: “to generate image embeddings in a semantically aware embedding space” [0045] 1st S) of the neural network model (via:
PNG
media_image25.png
820
874
media_image25.png
Greyscale
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu, KAZUHIRO does not teach the difference of claim 14 of:
a negative entropy.
Jaiswal teaches the difference of claim 14:
a negative entropy (“of discriminator predictions is minimized”, c.6,ll.30-35, “may be weighted with a multiplier a (tuned on 0.1, 1) in the overall objective”,c.6,ll. 45-50, “of the adversarial object-type discriminator 116”, c,7,ll. 60-65: fig. 1:116: “Adversarial Objective-type Discriminator”: fig.2: 105: “Class-agnostic object detector”).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu, KAZUHIRO teaches extraction, one of skill in the art of extraction can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Chiu, KAZUHIRO be as Jaiswal’s seeing the change “beneficial to downstream applications (e.g., application-specific object classification, visual search (object retrieval from large databases), computer vision-based speech processing, etc.) that can use such class-agnostic detections as inputs.”, Jaiswal, c.2,ll. 38-43.
Claim(s) 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over SALTZ et al. (US 2020/0388029 A1) in view of Wainrib et al. (US 2021/0256699 A1) as applied in claims 1,15,16,17,18,19 further in view of Chiu et al. (US 2020/0357143 A1) further in view of LIU et al. (CN 112163608 A) with SEARCH machine translation:
PNG
media_image26.png
906
910
media_image26.png
Greyscale
Claim 20 is rejected like claim 1,18,19:
Re 20. (Currently Amended), SALTZ of the combination (illustrated above) of SALTZ, Wainrib teaches A computer-implemented method for analyzing a whole slide image (WSI) in light of biological context, comprising:
extracting an embedding for each of a set of patches sampled from a WSI, wherein the embedding represents one or more histological features of the respective patch of the WSI;
for each of the patches of the set of patches, generating [[, by a neural network model,]] an encoded embedding by:
incorporating [[, by a first encoder,110]] a microscopic visual pattern related to the one or more histological features into the extracted embedding corresponding to the patch by attending to the embeddings of nearby patches in the set,
incorporating [[, by a second encoder,111 ]] a macroscopic visual pattern over the WSI as a whole into the extracted embedding corresponding to the patch by attending to the embeddings of all other patches (via “each patch”, SALTZ [0098], considered one at a time) in the set
generating a representation for the WSI by combining the encoded
performing a pathological task based on the representation for the WSI.
PNG
media_image27.png
1448
1129
media_image27.png
Greyscale
SALTZ of the combination (illustrated above) of SALTZ, Wainrib does not teach the difference of claim 20 of:
A) the embeddings)112…113
B)
E) macroscopic114…
C) the embeddings) …
D) visual patten) …
Chiu teaches the difference of claim 20 of:
A) rd S) (the embeddings)115…116
B) 117 on informative and stable image regions” [0031] 3rd S)
E) macroscopic118…
C) rd S) (the embeddings) …
D) nd S) visual patten).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib teaches feature extraction, one of skill in the art of feature extraction can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib be as Chiu’s seeing the change “improved visual localization”, Chiu [0010] 1st S:
PNG
media_image28.png
2087
1137
media_image28.png
Greyscale
SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu does not teach the last difference of claim 20 of:
E) macroscopic…
visual pattern).
LIU teaches the last difference of claim 20 of:
E) macroscopic…
119 information”, pg. 3, 3rd txt blk) (visual pattern).
Since SALTZ of the combination (illustrated above) of SALTZ, Wainrib,Chiu teaches identification, one of skill in the art can make SALTZ’s of the combination (illustrated above) of SALTZ, Wainrib,Chiu be as LIU’s seeing the change “to improve the accuracy of visual relationship identification”, LIU, pg. 2, 6th txt blk:
PNG
media_image29.png
2908
1137
media_image29.png
Greyscale
Conclusion
The prior art “nearest to the subject matter defined in the claims” (MPEP 707.05) made of record and not relied upon is considered pertinent to applicant's disclosure.
The following table lists several references that are relevant to the subject matter claimed and disclosed in this Application. The references are not relied on by the Examiner, but are provided to assist the Applicant in responding to this Office action.
Citation
Relevance
Barker et al. (US 2018/0374210 A1)
Barker discloses neighboring tiles:
[0081] 3rd S: “Tiles were adjacent120 to one another covering the entire tissue region of the slide.”
as the closest to the claimed “neighboring patches” of claim 1.
SHINAGAWA (US 2023/0090906 A1)
SHINAGAWA teaches local and global textures as patterns [0246] 3rd S:
Such image features may be generated by image analysis methods comprising the identification, analysis and/or measurement of objects, local and or global structures and/or textures121 present in any image data comprised in the first medical image IM1.
as the closest to the claimed “local pattern” and “global pattern” of claim 1.
Tellez et al. (Neural Image Compression for Gigapixel Histopathology Image Analysis)
Tellez teaches “global pattens…(e.g., tumor lesions)…and…local patterns (e.g., cells)” and “neighbor patches” and “Encoder” (fig.1) via:
page 568, 1.2 Neural Image Compression, 1st para, last S:
Finally, each embedding is placed into an array that keeps the original spatial arrangement intact so that neighbor embeddings in the array represent neighbor patches in the original image.
pg. 571, 4.1.1 Synthetic Dataset, 2nd para:
“To emulate global patterns in the images (e.g., tumor lesions), we defined two rectangles within each mask placed at random locations and characterized by their own orientation: one was either vertically or horizontally oriented (non-tilted); the other was tilted either 45 or 135 degrees (tilted). Each rectangle was associated to a randomly selected MNIST [30] digit class. To emulate local patterns (e.g., cells), instances of MNIST digits were placed through out the images at random locations. The class of these instances was determined by their spatial position, i.e., belonging to a certain rectangle class if placed within the boundaries of a rectangle or otherwise randomly selected. The label of each image was defined by the class of the tilted rectangle, with the non-tilted rectangle acting as a distraction. See Fig. 5 for an example image.”
PNG
media_image30.png
359
1053
media_image30.png
Greyscale
as the closest to the claimed “local pattern” and “global pattern…neighboring patches” of claim 1 and the “encoder” of claim 20.
Wang et al. (Label Cleaning Multiple Instance Learning: Refining Coarse Annotations on Single Whole-Slide Images)
Wang teaches a problem similar to applicant’s and multi-instance learning (“LC-MIL”) and incorporating global patterns (i.e., context) is a challenging problem and left to future work, page 3967, lcol, 2nd para:
Another limitation of LC-MIL, also shared by other binary classifiers, is that the model can be confused by irrelevant information when the histology structure is complex and heterogeneous. An example of this is represented by the difficulty in discriminating epithelial tissue from other types of tissues (loose connective tissue, smooth muscle), instead of focusing on learning the difference between benign and malignant tissue. On the other hand, this may not be a challenging task for a human (experienced) annotator. This could be addressed by extending the current version of LC-MIL to a multi-class setting. Moreover, pathologists usually make use of both local texture information and global information (e.g., shape122, location) to make decisions, while our proposed approach, as well as other patch-based methods, merely depend on the local texture and fail to incorporate global information. Efficiently incorporating this global context can be a challenging and significant problem, which we also leave as future work.
as the closest to applicant’s disclosed:
“without modeling the biological context of the patch with respect to other patches” and “global context” [3] last Ss:
Standard automated analysis techniques based on machine-learning methods take each patch as an independent unit without modeling the biological context of the patch with respect to other patches in the WSI during aggregation. This contrasts with diagnostic practice where a pathologist refers to both microscopic (regional or local within the WSI) patterns and macroscopic (global within the WSI) context when analyzing WSIs-multiple regions of the WSI are commonly picked as patterns of interest by the pathologist and evaluated as a whole in order to draw diagnostic and/or prognostic conclusions.
and as the closest to applicant’s disclosed “Multiple-instance learning”:
[32] Multiple-instance learning may be used to tackle both the need to break down WSIs into smaller image patches, as well as the issue with weak, slide-level labels. Multiple- instance learning operates at the patch level and identifies patches (e.g., regions) of a WSI that contribute to the weak label. WSIs are broken down into smaller patches and then, using a weak slide-level label only, a neural network is trained to identify which patches contribute to the slide-level label. Here, the main difficulty of multiple-instance learning lies in aggregating patch-level insights to the slide level.
KAPPEL (US 2022/0246244 A1)
KAPPEL teaches a “microscope” (fig. 8) and learning mapped tokens or “learn…token embeddings” as fig. 3:330: mapped tokens:
[0116] For example, the visual model is trained to predict token vectors as shown in FIG. 3. From a data repository 300 or a microscope during a running experiment images 310 may be passed as the independent variable to the input of a visual model 320. As the dependent variable the token embeddings 330, which have been mapped to the desired image classes, may be shown to the model at the output. The visual model may learn to predict token embeddings for each input.
PNG
media_image31.png
630
880
media_image31.png
Greyscale
as the closest to the claimed “slide” of claim 1 and “learnable token” in claims 11,12.
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DENNIS ROSARIO whose telephone number is (571)272-7397. The examiner can normally be reached Monday-Friday, 9AM-5PM EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Henok Shiferaw can be reached at 571-272-4637. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DENNIS ROSARIO/Examiner, Art Unit 2676
/MATTHEW C BELLA/Supervisory Patent Examiner, Art Unit 2667
1 extract: Mathematics. to determine (the root of a quantity that has a single root). (Dictionary.com)
2 embedding: Mathematics. the mapping of one set into another. (Dictionary.com)
3 encoding: to convert (a nerve signal) into a form that can be received by the brain (Dictionary.com)
4 representation: a mental image or idea so presented; concept. (Dictionary.com)
5 combine: to bring into or join in a close union or whole; unite, wherein union is defined: Mathematics. Also called join, logical sum, sum. the set consisting of elements each of which is in at least one of two or more given sets. ∪ (Dictionary.com)
6 task: a definite piece of work assigned to, falling to, or expected of a person; duty. (Dictionary.com)
7 MPEP 2132.04 All Claim Limitations Must Be Considered [R-01.2024], 3rd para. 2nd S:
Language (, by a neural network model,) that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
8 MPEP 2132.04 All Claim Limitations Must Be Considered [R-01.2024], 3rd para. 2nd S:
Language (, by a neural network model,) that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
9 machine learning: a branch of artificial intelligence in which a computer (claim 19’s “processor”) generates rules (i.e., claim 19’s “generate…an encoded embedding…by:”) underlying or based on raw data (claim 19’s “instructions”) that has been fed into it, wherein rules is defined: a prescribed method or procedure for solving a mathematical problem, or one constituting part of a computer program, usually expressed in an appropriate formalism, wherein instructions is defined: A sequence of bits that tells a computer's central processing unit to perform a particular operation. An instruction can also contain data to be used in the operation.(Dictionary.com)
10 MPEP 2106.07(a) II. WHEN MAKING A REJECTION, EXPLAIN WHY THE ADDITIONAL CLAIM ELEMENTS DO NOT RESULT IN THE CLAIM AS A WHOLE INTEGRATING THE JUDICIAL EXCEPTION INTO A PRACTICAL APPLICATION OR AMOUNTING TO SIGNIFICANTLY MORE THAN THE JUDICIAL EXCEPTION (STEP 2A PRONG TWO AND STEP 2B), last para:
In the event a rejection is made, it is a best practice for the examiner to consult the specification [33][34][35][62] to determine if there are elements ([62]:”the permutation-invariant transformer-based aggregation model”) that could be added to the claim to make it eligible. If so, the examiner should identify those elements in the Office action and suggest them as a way to overcome the rejection
11 Delete (same for claims 18,19,20) the commas such the claimed gerund (a noun) “generating” is adjectivally modified by the preposition modifier “by a neural network model”: --neural network model generating--
12 MPEP 2132.04 All Claim Limitations Must Be Considered [R-01.2024], 3rd para. 2nd S:
Language (, by a neural network model,) that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
13 global: comprehensive. (Dictionary.com)
14 macroscopic: pertaining to large units; comprehensive. (Dictionary.com)
15 histological: Biology. of or relating to organic tissues or their structure, wherein tissue is defined: Biology. an aggregate of similar cells and cell products forming a definite kind of structural material with a specific function, in a multicellular organism, wherein structure is defined: Biology. mode of organization; construction and arrangement of tissues, parts, or organs. (Dictionary.com): Examiner understands fig. 2A’s yellow-contour of 63/253,514 is a tumor (“that serves no function in the body”: Dictionary.com: CULTURAL: tumor) which is in contrast to histological that has a specific function.
16 cell: The basic unit of living matter in all organisms, consisting of protoplasm enclosed within a cell membrane. All cells except bacterial cells have a distinct nucleus that contains the cell's DNA as well as other structures (called organelles) that include mitochondria, the endoplasmic reticulum, and vacuoles. The main source of energy for all of a cell's biological processes is ATP. (Dictionary.com)
17 aberration: A deviation in the normal structure or number of chromosomes in an organism, wherein structure is defined: Biology. mode of organization; construction and arrangement of tissues, parts, or organs.(Dictionary.com)
18 decipher: to discover the meaning of (anything obscure or difficult to trace or understand). (Dictionary.com)
19 level: an extent, measure, or degree of intensity, achievement, etc.. (Dictionary.com)
20 fail: to be or become deficient or lacking; be insufficient or absent; fall short. (Dictionary.com)
21 information, wherein information is defined: knowledge communicated or received concerning a particular fact or circumstance; news, wherein fact is defined: that which actually exists or is the case; reality or truth, wherein truth is defined: the state or character of being true, wherein character is defined: the aggregate of features and traits that form the individual nature of some person or thing. (Dictionary.com)
22 histological ADJECTIVE
23 feature NOUN
24 form: Grammar .a word, part of a word, or group of words forming a construction that recurs in various contexts in a language with relatively constant meaning. (Dictionary.com)
25 histopathology: the science dealing with the histological structure of abnormal or diseased tissue; pathological histology. (Dictionary.com)
26 modifying as a participle, preposition, adjective (maybe as an adverb)
27 class: a number of persons or things regarded as forming a group by reason of common attributes, characteristics, qualities, or traits; kind; sort, wherein characteristic is defined: a distinguishing feature or quality. (Dictionary.com)
28 sub-: A prefix that means “underneath or lower” (as in subsoil), “a subordinate or secondary part of something else” (as in subphylum.), or “less than completely” (as in subtropical.) (Dictionary.com)
29 cellular: of, relating to, resembling, or composed of a cell or cells (Dictionary.com)
30 pattern: an example, instance, sample, or specimen. (Dictionary.com)
31 histological: Biology. of or relating to organic tissues or their structure, wherein tissue is defined: Biology. an aggregate of similar cells and cell products forming a definite kind of structural material with a specific function, in a multicellular organism. (Dictionary.com)
32 sub-: A prefix that means “underneath or lower” (as in subsoil), “a subordinate or secondary part of something else” (as in subphylum.), or “less than completely” (as in subtropical.) (Dictionary.com)
33 cellular: of, relating to, resembling, or composed of a cell or cells (Dictionary.com)
34 structure: Biology. mode of organization; construction and arrangement of tissues, parts, or organs, wherein organization is defined: organic structure; composition, wherein organic is defined: of, relating to, or affecting living tissue.(Dictionary.com)
35 histological: Biology. of or relating to organic tissues or their structure, wherein tissue is defined: Biology. an aggregate of similar cells and cell products forming a definite kind of structural material with a specific function, in a multicellular organism. (Dictionary.com)
36 histological: of or relating to histology or to the microscopic structure of the tissues of organisms, wherein histology is defined: 1: a branch of anatomy that deals with the minute structure of animal and plant tissues as discernible with the microscope; 2: tissue structure or organization (Meriam-Webster.com)
37 histological ADJECTIVE: Of or relating to histology, wherein histology NOUN is defined: The study of the structure, esp. the microscopic anatomy, of normal and abnormal tissues and organs of animal and plant bodies; a branch of biology or pathology dealing with this. Also: microscopic examination, or the microscopic appearance, of a tissue or structure (as shown in fig. [g] above), wherein Also is defined: Expressing amplification: as a further point, item, or circumstance tending in the same direction; further, in addition, besides, as well, too. (OED.com)
38 apparent: capable of being easily perceived or understood; plain or clear; obvious. (Dictionary.com)
39 dependencies: the state of being dependent; dependence. (Dictionary.com)
40 pattern: a distinctive style, model, or form. (Dictionary.com)
41 attention: the act or faculty of attending, especially by directing the mind to an object. (Dictionary.com)
42 possession: the act or fact of possessing, wherein posses is defined: to have knowledge of. (Dictionary.com)
43 dependencies: the state of being dependent; dependence. (Dictionary.com)
44 pattern: a distinctive style, model, or form. (Dictionary.com)
45 attention: the act or faculty of attending, especially by directing the mind to an object. (Dictionary.com)
46 scope: Linguistics, Logic. the range of words or elements (such as claim 1’s “extracting an embedding”) of an expression (claim 1) over which a modifier (a patent examiner) or operator (or me) has control. (Dictionary.com)
47 MATHEMATICAL CONCEPT: transformation: Also called transform. Logic. one of a set of algebraic formulas used to express the relations between elements, sets, etc., that form parts of a given system.
include: to contain, as a whole does parts or any part or element, wherein contain is defined: to be equal to (Dictionary.com).
48 MATHEMATICAL CONCEPT: function: Mathematics. a relationship in which an input value of a variable has a specifically calculated output value: for example, if the function of x is x 2 , the output will always be the square of whatever the value of x is. f, F (Dictionary.com)
49 MATHEMATICAL CONCEPT: transformation: Also called transform. Logic. one of a set of algebraic formulas used to express the relations between elements, sets, etc., that form parts of a given system.
50 include: to contain, as a whole does parts or any part or element, wherein contain is defined: to be equal to (Dictionary.com).
51 transformation: change in form, appearance, nature, or character. (Dictionary.com)
52 rotating: revolving around a central axis, line, or point (Dictionary.com)
53 MPEP 2111 Claim Interpretation; Broadest Reasonable Interpretation [R-10.2019] The broadest reasonable interpretation does not mean the broadest possible interpretation. Rather, the meaning given to a claim term [“transformations”] must be consistent with the ordinary and customary meaning of the term (unless the term has been given a special definition in the specification), and must be consistent with the use [The word “transformations” is used in a word-relationship, via the preposition “for”, to the disclosed word “rotating”] of the claim term in the specification [said Example 39: two paragraph specification: “These transformations can include affine transformations, for example, rotating”] and drawings [none in Example 39]. Further, the broadest reasonable interpretation of the claims must be consistent with the interpretation that those skilled in the art would reach. In re Cortright, 165 F.3d 1353, 1359, 49 USPQ2d 1464, 1468 (Fed. Cir. 1999)
54 transformation: the act or process of transforming. (Dictionary.com)
55 include: to contain, as a whole does parts or any part or element, wherein contain is defined: to be equal to (Dictionary.com).
56 rotating: revolving around a central axis, line, or point (Dictionary.com)
57 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
58 MPEP 2143.03 All Claim Limitations Must Be Considered [R-01.2024], 3rd S:
As a general matter, the grammar (non-restrictive phases: such as claim 1’s --, a neural network model,-- and ordinary meaning of terms as understood by one having ordinary skill in the art used in a claim will dictate whether, and to what extent, the language limits the claim scope. Language that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
59 Applicant’s disclosure [75]:
[75] At step 350, a downstream pathological task may be performed based on the representation for the WSI. The downstream pathological task may include, by way of example and not limitation, classifying the one or more histological features extracted from the WSI, classifying a pathological type of the WSI, predicting a progression risk of a disease associated with the one or more histological features, or determining a diagnosis of a patient associated with the WSI—
wherein may is defined: (used to express possibility). (Dictionary.com)
wherein diagnosis is defined: The identification by a medical provider of a condition, disease, or injury made by evaluating the symptoms and signs presented by a patient, wherein provider is defined: a person who supports a family or another person. (Dictionary.com)
60 task: a definite piece of work assigned to, falling to, or expected of a person; duty, wherein duty is defined: something that one is expected or required to do by moral or legal obligation, wherein legal is defined: permitted by law; lawful, wherein law is defined: any written or positive rule or collection of rules prescribed under the authority of the state or nation, as by the people in its constitution. (Dictionary.com) .
61 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
62 MPEP 2143.03 All Claim Limitations Must Be Considered [R-01.2024], 3rd S:
As a general matter, the grammar (non-restrictive phases: such as claim 1’s --, a neural network model,-- and ordinary meaning of terms as understood by one having ordinary skill in the art used in a claim will dictate whether, and to what extent, the language limits the claim scope. Language that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
63 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
64 neural network: Also called neural net. Computers. a hardware or software system in which weighted connections between data nodes are refined to produce increasingly accurate results in information processing, as in pattern recognition or problem solving, with the goal of algorithmic computing that requires minimal human intervention. (Dictionary.com)
65 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
66 MPEP 2143.03 All Claim Limitations Must Be Considered [R-01.2024], 3rd S:
As a general matter, the grammar (non-restrictive phases: such as claim 1’s --, a neural network model,-- and ordinary meaning of terms as understood by one having ordinary skill in the art used in a claim will dictate whether, and to what extent, the language limits the claim scope. Language that suggests or makes a feature or step optional but does not require that feature or step does not limit the scope of a claim under the broadest reasonable claim interpretation. In addition, when a claim requires selection of an element from a list of alternatives, the prior art teaches the element if one of the alternatives is taught by the prior art. See, e.g., Fresenius USA, Inc. v. Baxter Int’l, Inc., 582 F.3d 1288, 1298, 92 USPQ2d 1163, 1171 (Fed. Cir. 2009).
67 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
68 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
69 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
70 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
71
72 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
73 see Carmody’s claim 1, line 4,5’s “machine learning” that has no surrounding commas unlike claim 20’s -- , by a neural network model, --
74 machine learning: a branch of artificial intelligence in which a computer generates rules underlying or based on raw data that has been fed into it (Dictionary.com)
75 MPEP 2131 Anticipation — Application of 35 U.S.C. 102 [R-08.2017], 2nd param 2nd to last S:The elements must be arranged as required by the claim, but this is not an ipsissimis verbis test, i.e., identity of terminology is not required. In re Bond, 910 F.2d 831, 15 USPQ2d 1566 (Fed. Cir. 1990).
76 BROAD CLAIM LANGUAGE: “-ing” (of “comprising”): a suffix of nouns formed from verbs, expressing the action of the verb or its result, product, material, etc. (the art of building; a new building; cotton wadding )., wherein etc. is defined: and others; and so forth; and so on (used to indicate that more of the same sort or class might have been mentioned, but for brevity have been omitted), wherein so is defined: likewise or correspondingly; also; too. (Dictionary.com)
77 extracting: present participle of extract, wherein extract is defined: Mathematics.
a. to determine (the root of a quantity that has a single root).
b. to determine (a root of a quantity that has multiple roots). (Dictionary.com)
78 embedding: Mathematics. the mapping of one set into another. (Dictionary.com)
79 comma: the punctuation mark(,) indicating a slight pause in the spoken sentence and used where there is a listing of items or to separate a nonrestrictive clause or phrase from a main clause (Dictionary.com)
80 encode: to convert (a nerve signal) into a form that can be received by the brain (Dictionary.com)
81 encode: to convert (a nerve signal) into a form that can be received by the brain (Dictionary.com)
82 neural network:
1 an interconnected system of neurons, as in the brain or other parts of the nervous system
2 Also called: neural net. an analogous network of electronic components, esp one in a computer designed to mimic the operation of the human brain, wherein nervous is defined: of or relating to the nerves, wherein nerve is defined: Any of the bundles of fibers made up of neurons that carry sensory and motor information throughout the body in the form of electrical impulses., wherein sensory is defined: Involving the sense organs or the nerves that relay messages from them wherein relay is defined: Electricity. to retransmit (a signal, message, etc.) by or as if by means of a telegraphic relay, wherein telegraphic is defined: of or relating to the telegraph, wherein telegraph is defined: A communications system in which a message in the form of short, rapid electric impulses is sent, either by wire or radio, to a receiving station. Morse code is often used to encode messages in a form that is easily transmitted through electric impulses. (Dictionary.com)
83 BROAD CLAIM LANGUAGE: “-ing” (of “embedding”): a suffix of nouns formed from verbs (embed), expressing the action of the verb (embed) or its result (fig. 14A), product, material, etc. (the art of building; a new building; cotton wadding )., wherein etc. is defined: and others; and so forth; and so on (used to indicate that more of the same sort or class might have been mentioned, but for brevity have been omitted), wherein so is defined: likewise or correspondingly; also; too. (Dictionary.com)
84 BROAD CLAIM LANGUAGE (e.g., or the like): patch: a small piece, area, expanse, etc, wherein etc is defined: et cetera, wherein et cetera is defined: or the like; or something else similar (Dictionary.com)
85 image: a physical likeness or representation of a person, animal, or thing, photographed, painted, sculptured, or otherwise made visible. (Dictionary.com)
86 THE CLAIMED DIFFERENCE AS A WHOLE (regarding the claimed “neighboring” or the disclosed “nearby” applicant’s disclosure: [35] 2nd S):
the problem is in applicant’s disclosure [34] (“only dependencies across a selected subset of patches are modeled”:
[34] Certain techniques consider cross-patch dependencies during aggregation. One
example technique develops graph neural networks over pre-defined regions of interests as
vertices in order to capture significant regions of a slide. In another example technique, a single
distance layer is trained to measure the semantic similarity between a critical patch and other
patches in order to estimate the contribution of patches to the slide-level label. With these
techniques, only dependencies across a selected subset of patches are modeled, e.g.,
dependencies related to a single critical patch or over a pre-defined region.
The solution is the adding global data together with local data in fig. 2 (reproduced below: “+” sign boxed-in) as discussed in paragraph [65] (reproduced below), last S: “At all semantic encoders as in Figure 2(b ), the total self-attention ai.i of a patch i from a patch} may be a combination of spatial attention and semantic attention,”
87 BROAD CLAIM LANGUAGE (e.g., or the like): patch: a small piece, area, expanse, etc, wherein etc is defined: et cetera, wherein et cetera is defined: or the like; or something else similar (Dictionary.com)
88 (italics) represent claim limitations already taught
89 THE CLAIMED DIFFERENCE AS A WHOLE (regarding the claimed “neighboring” or the disclosed “nearby” applicant’s disclosure: [35] 2nd S):
the problem is in applicant’s disclosure [34] (“only dependencies across a selected subset of patches are modeled”:
[34] Certain techniques consider cross-patch dependencies during aggregation. One
example technique develops graph neural networks over pre-defined regions of interests as
vertices in order to capture significant regions of a slide. In another example technique, a single
distance layer is trained to measure the semantic similarity between a critical patch and other
patches in order to estimate the contribution of patches to the slide-level label. With these
techniques, only dependencies across a selected subset of patches are modeled, e.g.,
dependencies related to a single critical patch or over a pre-defined region.
The solution is the adding global data together with local data in fig. 2 (reproduced below: “+” sign boxed-in) as discussed in paragraph [65] (reproduced below), last S: “At all semantic encoders as in Figure 2(b ), the total self-attention ai.i of a patch i from a patch} may be a combination of spatial attention and semantic attention,”
90 (italics) represent claim limitations already taught
91 Applicant’s disclosure: [102] Herein, a computer-readable non-transitory storage medium or media may include one or more semiconductor-based or other integrated circuits (ICs) (such, as for example, field-programmable gate arrays (FPGAs) or application-specific ICs (ASICs)), hard disk drives (HDDs), hybrid hard drives (HHDs), optical discs, optical disc drives (ODDs), magneto-optical discs, magneto-optical drives, floppy diskettes, floppy disk drives (FDDs), magnetic tapes, solid-state drives (SSDs), RAM-drives, SECURE DIGITAL cards or drives, any other suitable computer-readable non-transitory storage media, or any suitable combination of two or more of these, where appropriate. A computer-readable non-transitory storage medium may be volatile, non-volatile, or a combination of volatile and non-volatile, where appropriate, wherein volatile is defined: Computers. of or relating to storage that does not retain data when electrical power is turned off or fails, wherein non-volatile is defined: (of computer memory) having the property of retaining data when electrical power fails or is turned off. (Dictionary.com)
92 typo: is interpreted as embedding
93 “clusters” is the object of “randomly selected”
94 “clusters” is the object of “randomly selected”
95 “clusters” is the object of “randomly selected”
96 stabilized: to make or hold stable, firm, or steadfast, wherein stable is defined: steadfast; not wavering or changeable, as in character or purpose; dependable, wherein dependable is defined: capable of being depended on; worthy of trust; reliable. (Dictionary.com)
97 (italics) represent claim limitations already taught
98 (italics) represent claim limitations already taught
99 each: every one of two or more considered individually or one by one, wherein every is defined: all possible; the greatest possible degree of, wherein one by one is defined: singly and successively, wherein singly is defined: apart from others; separately. (Dictionary.com)
100 BROAD CLAIM LANGUAGE: level: an extent, measure, or degree of intensity, achievement, etc.. (Dictionary.com)
101 BROAD CLAIM LANGUAGE: level: an extent, measure, or degree of intensity, achievement, etc.. (Dictionary.com)
102 BROAD CLAIM LANGUAGE: level: an extent, measure, or degree of intensity, achievement, etc.. (Dictionary.com)
103 transformer: Computers. Also transformer model a type of neural network that uses statistical relationships between sequential data, such as words and sentences, to learn context and produce output. (Dictionary.com)
104 MPEP 2141 II. Office Personnel As Factfinders, 1st para, 4th S:“In certain circumstances, it may also be important to include explicit findings as to how a person of ordinary skill would have understood prior art teachings, or what a person of ordinary skill would have known or could have done.”
105 image: a physical likeness or representation of a person, animal, or thing, photographed, painted, sculptured, or otherwise made visible. (Dictionary.com)
106 (italics) represent clam limitations already taught
107 ellipses (…) represent claim limitations already taught
108 (italics) represent clam limitations already taught
109 ellipses (…) represent claim limitations already taught
110 the non-restrictive phrase “, by a first encoder,” does not limit claim 20 under the broadest reasonable interpretation
111 the non-restrictive phrase “, by a second encoder,” does not limit claim 20 under the broadest reasonable interpretation
112 (italics) represents claim limitations already taught
113 ellipses (…) represent claim limitations already taught
114 Since SALTZ teaches “global” it is understood (see above examiner’s response to applicant’s remarks of the section III. PRIORITY regarding 35 USC 112(a) support for the claimed “global”) that SALTZ also teaches “macroscopic” since both “global” and “macroscopic” have the same meaning of “comprehensive” (Dictionary.com).
115 (italics) represents claim limitations already taught
116 ellipses (…) represent claim limitations already taught
117 focus: to be or become focused, where focus is defined: to bring to a focus or into focus; cause to converge on a perceived point, wherein focus is defined: a central point, as of attraction, attention, or activity. (Dictionary.com)
118 Note that the ipsissimis verbis test was unknowingly applied to “macroscopic” in the last Office action. Thus further mappings in addition to Saltz’s mapping to “macroscopic” “is not required” via MPEP 2131 Anticipation — Application of 35 U.S.C. 102 [R-08.2017], 2nd para, 2nd to last S: The elements must be arranged as required by the claim, but this is not an ipsissimis verbis test, i.e., identity of terminology is not required. In re Bond, 910 F.2d 831, 15 USPQ2d 1566 (Fed. Cir. 1990)
119 structure: mode of building, construction, or organization; arrangement of parts, elements, or constituents., wherein mode is defined: a particular type or form of something wherein form is defined: something that gives or determines shape; a mold, wherein shape is defined: something used to give form, as a mold or a pattern. (Dictionary.com)
120 adjacent: lying near, close, or contiguous; adjoining; neighboring. (Dictionary.com)
121 texture: the characteristic physical structure given to a material, an object, etc., by the size, shape, arrangement, and proportions of its parts, wherein shape is defined: something used to give form, as a mold or a pattern. (Dictionary.com)
122 shape: something used to give form, as a mold or a pattern. (Dictionary.com)
Prosecution Insights